Determining input impedance of tube preamp

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

If you put a 1000hz signal into the amp and have a series resistor when the
signal is 1/2 the value at the input the value of the the resistor is equal
to the actual input impedence.

Bob

"Crumb" <machine_nine@yahoo.com> wrote in message
news:7f446fbd.0409041528.3e6b6d99@posting.google.com...
> I have some RCA BC-2B mic preamps I am looking to pad the input with
> H-pads. My question is what values do I use for the Zin and Zout to
> calculate the H-pad?
>
> I asume that if I am using a 150 ohm mic, Zin would be 150 ohms and
> Zout would be the input impedance of the preamp... right?
>
> How do I go about calculating the input impedance of the preamp? Below
> is a copy of the RCA preamp schematic. Mine are stock and use the
> RCA/UTC transformers. The transformer's impedance ratio is 1:100.
> Sorry if this seems like a naive question.
>
>
> http://www.larryseyer.com/rcaschm.gif
>
> Thanks for your help.

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

"Crumb" <machine_nine@yahoo.com> wrote in message
news:7f446fbd.0409041528.3e6b6d99@posting.google.com...
> I have some RCA BC-2B mic preamps I am looking to pad the input with
> H-pads. My question is what values do I use for the Zin and Zout to
> calculate the H-pad?
>
> I asume that if I am using a 150 ohm mic, Zin would be 150 ohms and
> Zout would be the input impedance of the preamp... right?

No. Zin would be 1500 ohms or more to avoid excessive loading on the
microphone; most microphones are designed to be loaded by 10x their output
impedance. Ideally you design the pad so Zout is 150 ohms so the transformer
sees something like the source impedance it was intended to see, and given
the amount of attenuation you want, the resistor values fall into place. You
don't really need an H pad, just two series resistors (one on each leg)
followed by one shunt resistor. Microphone preamps don't work on an
impedance-matching basis, but on a bridging basis.

> How do I go about calculating the input impedance of the preamp? Below
> is a copy of the RCA preamp schematic. Mine are stock and use the
> RCA/UTC transformers. The transformer's impedance ratio is 1:100.
> Sorry if this seems like a naive question.

Bob already told you how to determine the input impedance of the preamp
using a series resistor. It's likely to be quite high, as the secondary of
the transformer is terminated only by the grid of the first tube stage,
which is very high impedance.

Didn't this quesiton get answered about a month ago?

Peace,
Paul

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

> Bob already told you how to determine the input impedance
> of the preamp using a series resistor. It's likely to be quite high,
> as the secondary of the transformer is terminated only by the grid
> of the first tube stage, which is very high impedance.

If this is correct (and I think it is), the primary is a very light load, so YOU
DON'T NEED TO MATCH IMPEDANCES.

Simply design your H pad for the output impedance of the mic, and everything
should be kop.

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

"Crumb" <machine_nine@yahoo.com

> I have some RCA BC-2B mic preamps ........



** Deja vu ???????






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

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

Selecting the wrong impedences, can alter your noise sensitivity and
frequency response.

Bob

"William Sommerwerck" <williams@nwlink.com> wrote in message
news:10jm7qojat39ob5@corp.supernews.com...
> > Bob already told you how to determine the input impedance
> > of the preamp using a series resistor. It's likely to be quite high,
> > as the secondary of the transformer is terminated only by the grid
> > of the first tube stage, which is very high impedance.
>
> If this is correct (and I think it is), the primary is a very light load,
so YOU
> DON'T NEED TO MATCH IMPEDANCES.
>
> Simply design your H pad for the output impedance of the mic, and
everything
> should be kop.
>
>

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

> Selecting the wrong impedences, can alter your
> noise sensitivity and frequency response.

Okay. So what is the impedance the tube itself "wants" to see?

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

A tube itself does not really care. The impedences the microphone sees and
if you are adding an attenuator what it sees and the cable lengths.

Bob

"William Sommerwerck" <williams@nwlink.com> wrote in message
news:10jmdfbdq6v52b4@corp.supernews.com...
> > Selecting the wrong impedences, can alter your
> > noise sensitivity and frequency response.
>
> Okay. So what is the impedance the tube itself "wants" to see?

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

"William Sommerwerck" <williams@nwlink.com> wrote in message
news:10jmdfbdq6v52b4@corp.supernews.com...
> > Selecting the wrong impedences, can alter your
> > noise sensitivity and frequency response.
>
> Okay. So what is the impedance the tube itself "wants" to see?

As Bob says, the tube doesn't really care. But it will give its best noise
performance when the source resistance feeding the grid is at least 3.85x
the total equivalent noise resistance of the tube and its associated
components. (For this exercise I define "best noise performance" as adding
less than 1dB of noise to the inherent noise of the microphone, which is
assumed to be a 150-ohm resistor.)

That total noise resistance has two or three components:

1) The inherent noise resistance of the tube. In a triode, this
(oversimplifying) is typically 2.5/gm, where gm is the transconductance of
the tube at its actual operating point, in mhos -- excuse me, Siemenses.

2) The plate resistor, divided by the square of the tube's gain.

3) The cathode resistor, if unbypassed.

Add all these resistances up, and you get the equivalent noise resistance.
So for a tube with a gain of 10x, a plate resistor of 50k, a cathode
resistor of 2.2k, unbypassed, and a gm of 2 milliSiemens, you get an
equivalent noise resistance of:

[50k/100] + [2.5/0.002] + 2200 = 500 +1250 + 2200 = 3950.

3.85 x 3950 = 15,207; in practice, a transformer with a secondary of 15k or
higher will do fine. The LS-series transformer used in the RCA preamp under
discussion has such a secondary; so does the Jensen JE-115K-E, widely used
in tube preamp designs.

Peace,
Paul

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

> As Bob says, the tube doesn't really care. But it will give its best
> noise performance when the source resistance feeding the grid is
> at least 3.85x the total equivalent noise resistance of the tube
> and its associated components.

Then it DEOS "care."

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

On Sun, 5 Sep 2004 08:49:14 -0700, "William Sommerwerck"
<williams@nwlink.com> wrote:

>Okay. So what is the impedance the tube itself "wants" to see?

The tube is noise matched at impossibly high impedances,
unreachable with real world transformers of adequate
bandwidth.

Source impedance to the transformer will affect *its*
frequency response, and generally, lower is better.

Chris Hornbeck

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

With op amps most noise is generated in the input resistors due to the
current through them and their temperature.

B0b

"William Sommerwerck" <williams@nwlink.com> wrote in message
news:10jn0vt1f3aal52@corp.supernews.com...
> > As Bob says, the tube doesn't really care. But it will give its best
> > noise performance when the source resistance feeding the grid is
> > at least 3.85x the total equivalent noise resistance of the tube
> > and its associated components.
>
> Then it DEOS "care."

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

> With op amps most noise is generated in the input resistors
> due to the current through them and their temperature.

The current through the input resistors (in a passive attenuator) is minuscule.
The "excess noise" generated is of no concern.

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

Most transformer specs will tell you the swamping resistor value for optimum
performance of that particular transformer. The resistor lowers the Q of the
transformer to flatten out its resonant peak. The voltage gain of the
transformer increases the input signal so that you increase the signal level
higher above the noise floor. the less gain you need the less you amplify
the input noise.

Bob

"Chris Hornbeck" <chrishornbeckremovethis@att.net> wrote in message
news:2bimj0ldtggqvjogkgjpa8puhqldosftbt@4ax.com...
> On Sun, 5 Sep 2004 08:49:14 -0700, "William Sommerwerck"
> <williams@nwlink.com> wrote:
>
> >Okay. So what is the impedance the tube itself "wants" to see?
>
> The tube is noise matched at impossibly high impedances,
> unreachable with real world transformers of adequate
> bandwidth.
>
> Source impedance to the transformer will affect *its*
> frequency response, and generally, lower is better.
>
> Chris Hornbeck

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

The least noise is with a microphone working on the principle of a laser
interferometer. The only references I have seen to that type of microphone
for sale is for pointing at distant windows to spy.

Bob

"sycochkn" <sycochkn@earthlink.net> wrote in message
news:gHL_c.1082$ip2.554@newsread3.news.pas.earthlink.net...
> Most transformer specs will tell you the swamping resistor value for
optimum
> performance of that particular transformer. The resistor lowers the Q of
the
> transformer to flatten out its resonant peak. The voltage gain of the
> transformer increases the input signal so that you increase the signal
level
> higher above the noise floor. the less gain you need the less you amplify
> the input noise.
>
> Bob
>
> "Chris Hornbeck" <chrishornbeckremovethis@att.net> wrote in message
> news:2bimj0ldtggqvjogkgjpa8puhqldosftbt@4ax.com...
> > On Sun, 5 Sep 2004 08:49:14 -0700, "William Sommerwerck"
> > <williams@nwlink.com> wrote:
> >
> > >Okay. So what is the impedance the tube itself "wants" to see?
> >
> > The tube is noise matched at impossibly high impedances,
> > unreachable with real world transformers of adequate
> > bandwidth.
> >
> > Source impedance to the transformer will affect *its*
> > frequency response, and generally, lower is better.
> >
> > Chris Hornbeck
>
>

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

sycochkn wrote:

> With op amps most noise is generated in the input resistors

Such a design would be a very poor example.

> due to the current through them

You're suggesting that input current noise density ( as opposed to
voltage
) is a big factor ?

> and their temperature.

Does thermal noise go away if you're not using op-amps ?

What kind of circuit toplogy are you thinking of ?

Graham

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

>The tube is noise matched at impossibly high impedances,
>unreachable with real world transformers of adequate
>bandwidth.

On reflection, this is a very poor model for audio use and
should be ignored. Paul's noise figure calculations tell us
what we really need to know, and eliminate the Ghafla.

Bob's mention of very large area microphones is quite
intriguing.

Chris Hornbeck

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

"William Sommerwerck" <williams@nwlink.com> wrote in message
news:10jn0vt1f3aal52@corp.supernews.com...
> > As Bob says, the tube doesn't really care. But it will give its best
> > noise performance when the source resistance feeding the grid is
> > at least 3.85x the total equivalent noise resistance of the tube
> > and its associated components.
>
> Then it DOES "care."

Not in the same way that, for example, a bipolar transistor does. The tube
and its associated components will add a particular amount of noise to the
input signal regardless of what the input impedance may be. (Whether or not
that particular amount of noise is significant was the point of my little
exercise.) A bipolar transistor, on the other hand, in addition to noise
voltage, produces noise current, which produces a noise voltage when it's
run through an input impedance. So the transistor is interacting with its
source impedance and so could be said to "care" about what impedance feeds
it, and in fact there's an optimum impedance at which the transistor (or
circuit with several transistors, e.g. an op amp) will add the least noise.
Whereas the tube only makes its own noise, and doesn't stimulate the source
to produce more. Hmm, I think I'm personifying too much here.

Anyhow, it's a pretty fine semantic point, but it does illustrate a
real-world difference in the behavior of tubes and bipolar transistors.
(FETs act more like tubes in this regard.)

Peace,
Paul

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

"William Sommerwerck" <williams@nwlink.com> wrote in message
news:10jn46q3ldqhi0c@corp.supernews.com...
> > With op amps most noise is generated in the input resistors
> > due to the current through them and their temperature.
>
> The current through the input resistors (in a passive attenuator) is
minuscule.
> The "excess noise" generated is of no concern.

I think he's referring here to noise current, which is generated by
bipolar-input op amps but not (in any significant amount) by FET-input
ditto. But we're getting off-topic, a little.

Peace,
Paul

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

"sycochkn"
>
> With op amps most noise is generated in the input resistors due to the
> current through them and their temperature.
>


** Bullshit !!

What dickhead's web site did you get that idiot idea from ?




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

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

> ** Bullshit !!

Like a trout on a lure....

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

In article <2q1tmmFp9321U1@uni-berlin.de> philallison@tpg.com.au writes:

> ** Bullshit !!
>
> What dickhead's web site did you get that idiot idea from ?

This is the reason why I often aske people not to point me to a web
site, but to explain a statement in their own words. A too-short
answer is ambiguous. A more complete explanation helps me to see what
the poster is actually thinking about, which may not be exactly what
he wrote but is a useful piece of the puzzle.

The shorter an answer, the closer it is to being either completely
right or completely wrong. 50% chance of being right (though
possbly irrelevant to the subject) seems to be the preferred
option for some people here.

--
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

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

"sycochkn"
> The least noise is with a microphone working on the principle of a laser
> interferometer. The only references I have seen to that type of microphone
> for sale is for pointing at distant windows to spy.
>


** I would expect a "microphone" with a diaphragm area of several square
metres to have low noise.




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

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

"Paul Stamler"
>
> Not in the same way that, for example, a bipolar transistor does. The tube
> and its associated components will add a particular amount of noise to the
> input signal regardless of what the input impedance may be. (Whether or
not
> that particular amount of noise is significant was the point of my little
> exercise.) A bipolar transistor, on the other hand, in addition to noise
> voltage, produces noise current, which produces a noise voltage when it's
> run through an input impedance. So the transistor is interacting with its
> source impedance and so could be said to "care" about what impedance feeds
> it, and in fact there's an optimum impedance at which the transistor (or
> circuit with several transistors, e.g. an op amp) will add the least
noise.
>


** WRONG - in both cases noise is least with a low source impedance.

An input is at its quietest with a short applied.

You have mixed up "noise figure" with actual noise level.






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

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

> An input is at its quietest with a short applied.

That's now what I was taught in school.

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

On Mon, 6 Sep 2004 03:43:19 -0700, "William Sommerwerck"
<williams@nwlink.com> wrote:

>> An input is at its quietest with a short applied.
>
>That's now what I was taught in school.

It's correct, but as usual irrelevant. We were discussing
the case of an input transformer giving nominal choices of
source impedance.

Each active device has an optimum input transformer ratio
to provide its optimum source impedance from a given external
impedance.

Chris Hornbeck

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

In article <mDL_c.8252$w%6.5962@newsread1.news.pas.earthlink.net>,
"sycochkn" <sycochkn@earthlink.net> wrote:
> With op amps most noise is generated in the input resistors due to the
> current through them and their temperature.

In most mike amps, the input resistors are large and in parallel with
the source (the mike itself) and thus affect the noise performance very
little. If anything, they serve to lower the effective resistance that
the input stage sees.

To the OP: a mike amp is a bridging load, so it's not a huge source of
error to design your pad as if it were plugged into an open circuit.

Make the impedance between pin 2 and 3 of the pad's output equal to 150
ohms and you will not adversely affect the noise performance of the mike
amp. Similarly, make sure to include the mike's nominal 150 ohm source
impedance in your calculations.


Regards,

Monte McGuire
monte.mcguire@verizon.net

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

Crumb wrote:

> I have some RCA BC-2B mic preamps I am looking to pad the input with
> H-pads. My question is what values do I use for the Zin and Zout to
> calculate the H-pad?
>
> I asume that if I am using a 150 ohm mic, Zin would be 150 ohms and
> Zout would be the input impedance of the preamp... right?
>
> How do I go about calculating the input impedance of the preamp? Below
> is a copy of the RCA preamp schematic. Mine are stock and use the
> RCA/UTC transformers. The transformer's impedance ratio is 1:100.
> Sorry if this seems like a naive question.
>
> http://www.larryseyer.com/rcaschm.gif

I just realised that the above circuit uses grid leak bias !

Can anyone explain a good reason for this ?


Graham

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

"Pooh Bear" <rabbitsfriendsandrelations@hotmail.com> wrote in message
news:413D2C61.13968721@hotmail.com...

> > http://www.larryseyer.com/rcaschm.gif
>
> I just realised that the above circuit uses grid leak bias !
>
> Can anyone explain a good reason for this ?

It doesn't really. While there's no resistor connecting the first tube's
grid to ground, the transformer's secondary has a fairly low DC resistance.
The 8.2k resistor from its cathode to ground sets the bias point in a
standard self-bias setup, albeit with an unusually high value for the
cathode resistor. High enough to increase the preamp's noise figure, in
fact.

Peace,
Paul

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

"Pooh Bear"

> > http://www.larryseyer.com/rcaschm.gif
>
> I just realised that the above circuit uses grid leak bias !
>


** Better lay off all those bee hives Pooh.

There are 8.2 k and 1 k cathode resistors for bias.




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

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

Paul Stamler wrote:

> "Pooh Bear" <rabbitsfriendsandrelations@hotmail.com> wrote in message
> news:413D2C61.13968721@hotmail.com...
>
> > > http://www.larryseyer.com/rcaschm.gif
> >
> > I just realised that the above circuit uses grid leak bias !
> >
> > Can anyone explain a good reason for this ?
>
> It doesn't really. While there's no resistor connecting the first tube's
> grid to ground, the transformer's secondary has a fairly low DC resistance.

Yup, I had a bit of a 'blonde moment' there.


> The 8.2k resistor from its cathode to ground sets the bias point in a
> standard self-bias setup, albeit with an unusually high value for the
> cathode resistor. High enough to increase the preamp's noise figure, in
> fact.

That's true.


Graham

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

I usually use a 12AX7 with 1k cathode resistor and 100k plate resistor and
100k resistor to ground from the grid.
I would probably also have a swamping resistor across the transformer and
couple to the grid resistor through a capacitor. a 100 to 1 impedence ratio
infers a 10 to 1 voltage or turns ratio ratio. is the feed back circuit
essential. So there is a second stage to go back to low impedence balanced
output.

Bob

"Pooh Bear" <rabbitsfriendsandrelations@hotmail.com> wrote in message
news:413DBE96.48AD0BB@hotmail.com...
> Paul Stamler wrote:
>
> > "Pooh Bear" <rabbitsfriendsandrelations@hotmail.com> wrote in message
> > news:413D2C61.13968721@hotmail.com...
> >
> > > > http://www.larryseyer.com/rcaschm.gif
> > >
> > > I just realised that the above circuit uses grid leak bias !
> > >
> > > Can anyone explain a good reason for this ?
> >
> > It doesn't really. While there's no resistor connecting the first tube's
> > grid to ground, the transformer's secondary has a fairly low DC
resistance.
>
> Yup, I had a bit of a 'blonde moment' there.
>
>
> > The 8.2k resistor from its cathode to ground sets the bias point in a
> > standard self-bias setup, albeit with an unusually high value for the
> > cathode resistor. High enough to increase the preamp's noise figure, in
> > fact.
>
> That's true.
>
>
> Graham
>
>

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

"sycochkn" <sycochkn@earthlink.net> wrote in message
news 1s%c.744$xA1.1@newsread3.news.pas.earthlink.net...
> I usually use a 12AX7 with 1k cathode resistor and 100k plate resistor and
> 100k resistor to ground from the grid.
> I would probably also have a swamping resistor across the transformer and
> couple to the grid resistor through a capacitor. a 100 to 1 impedence
ratio
> infers a 10 to 1 voltage or turns ratio ratio. is the feed back circuit
> essential. So there is a second stage to go back to low impedence balanced
> output.

Your practice sounds good except for the resistor to ground from the grid.
If you're using a 1:10 turns ratio transformer, its secondary impedance will
be 15k, assuming a standard 150 ohm microphone, and you'd normally load that
with a 150k resistor. With many transformers 100k will overdamp it at high
frequencies.

Is the feedback circuit essential? Yes, in this design, for a couple of
reasons. First off, it lowers the distortion of the circuit (that's a whole
'nother discussion). Second, it lowers the gain, which is good, because the
two medium-to-high mu tubes in sequence would give you way too much gain on
their own. As is the circuit has something like 55dB of gain to the plate of
the output tube (okay, a little less, but not much less), which is why the
original poster was looking for pad circuits. Without the feedback network
you'd probably see something like 70dB.

Peace,
Paul