transformer impedance

[Walter Harley]

Audio transformers are parameterized by their impedance, among other things.
For instance, one can buy a 600 ohm 1:1 transformer, for isolating a nominal
600 ohm "pro audio" line level signal, or a 10k 1:1 transformer for doing
the same to a higher-impedance "consumer level" signal.

In thinking about it, I realize that I really don't understand what this
impedance is, nor how to measure it, nor how to work with it in a circuit.

Are there any good explanations online, or in commonly available texts? I
just poked around in AoE and didn't find any details, but maybe I missed it.
The ARRL Handbook says "The primary terminal impedance of an iron-core
transfomer is determined wholly by the load connected to the secondary and
by the turns ratio."

What does it actually mean, for a transformer winding to have (say) a 600
ohm impedance?

Given an unknown transformer, how would I measure its nominal impedance?

In a typical audio situation, source impedance might be 300 ohms, load
impedance 10k. Assuming one wants a 1:1 voltage transfer, what would be the
appropriate transformer impedance? What would be the consequences of
selecting a transformer with the wrong nominal impedance?

Thanks for any help you all can give me in understanding this topic!

-walter

 

[John Popelish]

Walter Harley wrote:

Audio transformers are parameterized by their impedance, among other things.
For instance, one can buy a 600 ohm 1:1 transformer, for isolating a nominal
600 ohm "pro audio" line level signal, or a 10k 1:1 transformer for doing
the same to a higher-impedance "consumer level" signal.

In thinking about it, I realize that I really don't understand what this
impedance is, nor how to measure it, nor how to work with it in a circuit.

Are there any good explanations online, or in commonly available texts? I
just poked around in AoE and didn't find any details, but maybe I missed it.
The ARRL Handbook says "The primary terminal impedance of an iron-core
transfomer is determined wholly by the load connected to the secondary and
by the turns ratio."

What does it actually mean, for a transformer winding to have (say) a 600
ohm impedance?

Given an unknown transformer, how would I measure its nominal impedance?

In a typical audio situation, source impedance might be 300 ohms, load
impedance 10k. Assuming one wants a 1:1 voltage transfer, what would be the
appropriate transformer impedance? What would be the consequences of
selecting a transformer with the wrong nominal impedance?

The specified impedance is not the impedance of the transformer,
itself, but the source and load impedance that makes the transformer
operate most ideally over the specified frequency range. For example,
the 600 ohm transformer approaches an open load inductive impedance of
about 600 ohms at the lower end of the specified frequency range, and
the leakage inductance approaches 600 ohms at the upper end of that
range. Between those extremes, the transformer is essentially
invisible (shunt inductance parallels an acceptable impedance much
higher than the load, and the leakage inductance in series consumes an
acceptable fraction of the total signal voltage). If you use a
different frequency range, the optimum source and load impedances
would shift.

 

[Tim Williams]

"Walter Harley" <walterh@cafewalterNOSPAM.com> wrote in message
news:aYidnV5w2eSxTZ7eRVn-vA@speakeasy.net...

What does it actually mean, for a transformer winding to have (say)
a 600 ohm impedance?

Just best all around parameters. The only things that seperate a real
transformer from an ideal transformer are resistance, inductance and
capacitance. As you might've guessed from the pattern of wires crossing,
it's very complex but can be approximated with three lumped constants: DC
resistance of the windings, inductance of the winding in question and bulk
capacitance. From there, you specify parameters to meet specifications,
namely loss and frequency range. A transformer with 5% loss from 20Hz to
10kHz at 1kohm is the exact same as 10Hz to 20kHz, at 500 ohms. The loss
doubles to 10% because the DCR remains constant but you are using it at half
the impedance.

When you add a magnetic core, you add three more parameters: saturation at
low frequency and hysteresis and eddy current losses at high frequency.
Saturation acts similar to inductance but worse: instead of mere reduced
impedance at low frequency, you get a reduced *inductance* at a certain
field strength. (Gapping the core reduces the "core-ness" (i.e., reduced mu
and nonlinearity) of the core by adding some air core back in...(kinda like
using a small amount of negative feedback in an amplifier, where gain and
distortion are reduced)). This stabs an absolute power x frequency limit on
the transformer.

Given an unknown transformer, how would I measure its nominal impedance?

....So to answer your question, I have no idea! It simply depends too much.

In a typical audio situation, source impedance might be 300 ohms, load
impedance 10k. Assuming one wants a 1:1 voltage transfer, what would
be the appropriate transformer impedance? What would be the
consequences of selecting a transformer with the wrong nominal
impedance?

The driving signal can handle a 300 ohm transformer so that would be the
best option. It'll also drive the 10k load solidly (good "regulation" you
could say). But since the load is "only" 10k, you could use a 10k:10k
transformer and accept a little more loss (the 300 ohm source would rather
put more current through that primary!) for extended frequency response
(although both transformers should be wound so frequency doesn't come into
it too much, keep this in mind).

The best of both worlds, particularly if you are transferring power, is to
match source to load. But you knew that! ;-)

Tim

--
Deep Fryer: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

 

[John Woodgate]

I read in sci.electronics.design that Walter Harley
<walterh@cafewalterNOSPAM.com> wrote (in
<aYidnV5w2eSxTZ7eRVn-vA@speakeasy.net&gt about 'transformer impedance',
on Wed, 17 Aug 2005:

Audio transformers are parameterized by their impedance, among other
things. For instance, one can buy a 600 ohm 1:1 transformer, for
isolating a nominal 600 ohm "pro audio" line level signal, or a 10k 1:1
transformer for doing the same to a higher-impedance "consumer level"
signal.

In thinking about it, I realize that I really don't understand what
this impedance is, nor how to measure it, nor how to work with it in a
circuit.

It's a (somewhat convenient, but, as you've found, confusing) fiction,
like loudspeaker 'power' ratings.

Are there any good explanations online, or in commonly available texts?
I just poked around in AoE and didn't find any details, but maybe I
missed it. The ARRL Handbook says "The primary terminal impedance of an
iron-core transfomer is determined wholly by the load connected to the
secondary and by the turns ratio."

That is correct, if winding resistances are not too high for proper
operation. Keep it in mind.

What does it actually mean, for a transformer winding to have (say) a
600 ohm impedance?

It means it's

***designed for use with a 600 ohm load, in circuits with the voltage
levels associated with '600 ohm technology', i.e. from a few hundred
millivolts to perhaps 20 V (+22 dBu)***.

The impedance isn't an electrical property of a winding or of the
transformer as a whole.

Given an unknown transformer, how would I measure its nominal
impedance?

You can't, because it's not a property of the transformer. You can only
measure its properties and deduce what sort of circuits it's intended
for, For example, if a winding has a resistance of 300 ohms, it's likely
to be for '10 kohm' circuits than 600 ohm. If you apply 0 dBu( 0.775 V)
to a winding and find that the transformer is saturated (output waveform
distorted), it's probably a transformer for a microphone input of a
mixer.

In a typical audio situation, source impedance might be 300 ohms,

That's a bit high these days; 50 to 75 ohms is more usual.

load impedance 10k. Assuming one wants a 1:1 voltage transfer, what
would be the appropriate transformer impedance?

1:1 voltage ratio means 1:1 turns ratio, if losses are negligible. Since
you have a low-impedance source and a high impedance load, winding
resistance losses are indeed minimal, so a precise 1:1 turns ratio will
do.

What would be the consequences of selecting a transformer with the
wrong nominal impedance?

It's not possible to tell, because what you call the 'nominal impedance'
is actually irrelevant. If the winding resistances are low enough, and
the transformer will accept the maximum voltage you want to apply
without saturating, it will work.
--
Regards, John Woodgate, OOO - Own Opinions Only.
There are two sides to every question, except
'What is a Moebius strip?'
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

 

[John Woodgate]

I read in sci.electronics.design that John Popelish <jpopelish@rica.net>
wrote (in <8pednZ2dnZ1gbrG4nZ2dnQZOnt6dnZ2dRVn-zp2dnZ0@adelphia.com&gt
about 'transformer impedance', on Wed, 17 Aug 2005:

The specified impedance is not the impedance of the transformer,
itself, but the source and load impedance that makes the transformer
operate most ideally over the specified frequency range.

All OK so far.

For example, the 600 ohm transformer approaches an open load
inductive impedance of about 600 ohms at the lower end of the specified
frequency range, and the leakage inductance approaches 600 ohms at the
upper end of that range.

Only for crap transformers from Sum Yung Gai Inc. For decent
transformers, the frequency range is specified for VERY SMALL values of
input impedance reduction at the low frequency and insertion loss at the
high frequency end.

Between those extremes, the transformer is essentially invisible (shunt
inductance parallels an acceptable impedance much higher than the load,
and the leakage inductance in series consumes an acceptable fraction of
the total signal voltage). If you use a different frequency range, the
optimum source and load impedances would shift.

Not really: any reasonable 'different' frequency range would overlap the
specified range to a large extent (e.g. spec range 50 Hz to 16 kHz,
'different' range 40 Hz to 10 kHz), and the optimum impedances wouldn't
change significantly. The winding resistances are involved, as well as
the primary shunt and leakage inductances.
--
Regards, John Woodgate, OOO - Own Opinions Only.
There are two sides to every question, except
'What is a Moebius strip?'
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

 

[John Woodgate]

I read in sci.electronics.design that Tim Williams
<tmoranwms@charter.net> wrote (in <1ARMe.6900$rc6.6234@fe03.lga&gt about
'transformer impedance', on Wed, 17 Aug 2005:

The best of both worlds, particularly if you are transferring power, is
to match source to load. But you knew that! ;-)

NOT FOR AUDIO! Matching for maximum power transfer is never used in
normal audio technology, for good and sufficient reasons. The rule is
'low-impedance source, high impedance load'.
--
Regards, John Woodgate, OOO - Own Opinions Only.
There are two sides to every question, except
'What is a Moebius strip?'
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

 

[Ban]

Walter Harley wrote:

Audio transformers are parameterized by their impedance, among other
things. For instance, one can buy a 600 ohm 1:1 transformer, for
isolating a nominal 600 ohm "pro audio" line level signal, or a 10k
1:1 transformer for doing the same to a higher-impedance "consumer
level" signal.

Nope, now the pro audio is defined in dBu, which means "unspecified"
impedance. The output might range up to 500R and the input impedance is
usually greater than 10k. Microphone inputs are balanced with more than 4k
impedance. Here 600R would result in unspecified behaviour of the mike.

In thinking about it, I realize that I really don't understand what
this impedance is, nor how to measure it, nor how to work with it in
a circuit.

It will tell you how much current the output stage has to provide. If the
voltage is +/-12Vp, the current in 600R would be +/-20mApeak. into 10k the
current would be +/-1.2mA only.

Are there any good explanations online, or in commonly available
texts? I just poked around in AoE and didn't find any details, but
maybe I missed it. The ARRL Handbook says "The primary terminal
impedance of an iron-core transfomer is determined wholly by the load
connected to the secondary and by the turns ratio."

What does it actually mean, for a transformer winding to have (say) a
600 ohm impedance?

Given an unknown transformer, how would I measure its nominal
impedance?

Transformers are hardly used any more, because of cost, size and weight. You
can not easily measure its intended working impedance. What you can measure
is its DC-resistance. It has to be *much* smaller, to not cause too much
insertion loss.

In a typical audio situation, source impedance might be 300 ohms, load
impedance 10k. Assuming one wants a 1:1 voltage transfer, what would
be the appropriate transformer impedance? What would be the
consequences of selecting a transformer with the wrong nominal
impedance?

The source impedance doesn't matter as long as it is small. A 1:1
Transformer has equal impedances on both sides (exept the losses), so if you
connect the secondary you will measure 10k on the primary as well. The
transformer itself has no impedance, but is intended to work with the
specified impedance. If there is a big mismatch, the frequency response
might suffer and/or saturating(distortion!) can occur, especially at very
low frequencies.

--
ciao Ban
Bordighera, Italy

 

[Ban]

Tim Williams wrote:

The best of both worlds, particularly if you are transferring power,
is to match source to load. But you knew that! ;-)

Nope, this is not recommended. It will reduce the level by 6dB each stage
and waste a lot of heat unnecessarily. The same is true for audio
amplifiers. Their output impedance should approach 0, otherwise the
loudspeakers might develop additional errors in the frequency response.

--
ciao Ban
Bordighera, Italy

 

[Tim Williams]

"Ban" <bansuri@web.de> wrote in message
newsxXMe.60879$2U1.3305187@news3.tin.it...

The best of both worlds, particularly if you are transferring power,
is to match source to load. But you knew that! ;-)

Nope, this is not recommended. It will reduce the level by 6dB each
stage and waste a lot of heat unnecessarily. The same is true for audio
amplifiers. Their output impedance should approach 0

Huh? Oh, you're talking Zo, not "best power output" impedance. I was
talking load matching, obviously.

Tim

--
Deep Fryer: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

 

[Joerg]

Hello Ban,

Transformers are hardly used any more, because of cost, size and weight. You
can not easily measure its intended working impedance. What you can measure
is its DC-resistance. It has to be *much* smaller, to not cause too much
insertion loss.

That is what someone else in this forum also said (Martin?). But
transformers are still used when there is no other viable option. These
are mostly hardcore EMI and other interference situations.

Regards, Joerg

http://www.analogconsultants.com

 

[Walter Harley]

"Joerg" <notthisjoergsch@removethispacbell.net> wrote in message
newsJvNe.2687$AT7.444@newssvr22.news.prodigy.net...

That is what someone else in this forum also said (Martin?). But
transformers are still used when there is no other viable option. These
are mostly hardcore EMI and other interference situations.

Transformers are still widely used in pro audio - perhaps not in the
"pro-sumer" category of gear, but in quite a lot of higher-end gear; just
take a quick look through the ads in any pro audio magazine. This is partly
because of audio engineers' habit of running millivolt-level signals through
hundreds of feet of cable and still expecting S/N ratios of 80dB or better,
and partly because of the mystique of successful recordings made with
transformer gear in the past.

The problems with transformers, certainly, are that they are expensive and
large. But the common-mode rejection, ESD and EMI handling, distortion, and
noise specs of a good audio transformer are really quite hard to beat. (Bad
ones, on the other hand, are lousy.)

 

[martin griffith]

On Sat, 20 Aug 2005 01:31:27 GMT, in sci.electronics.design Joerg
<notthisjoergsch@removethispacbell.net> wrote:

Hello Ban,

Transformers are hardly used any more, because of cost, size and weight. You
can not easily measure its intended working impedance. What you can measure
is its DC-resistance. It has to be *much* smaller, to not cause too much
insertion loss.

That is what someone else in this forum also said (Martin?). But
transformers are still used when there is no other viable option. These
are mostly hardcore EMI and other interference situations.

Regards, Joerg

http://www.analogconsultants.com
I dont thik it was me. But generally I agree. There was a thread many

months ago about relays, me Vs Win, more or less. Cant remember who
"won" though. Probably Win

For audio a good transformer , say Jensen JE16 or Sowter equivalent,
in conjunction with star quad cable is incredible robust in terms of
CMRR. generally for mic levels. The bog standard single or triple
opamp circuitis usually adequate for line level signals.

One test I have always meant to do is a CM 20KHz to100MHz AM, 50% mod.
sweep of opamps micamps (INA103 etc)to see how good they are under
difficult conditions, ie 60dB gain. This AFAIK is never published in
the spec sheets

</waffle>


martin

 

[John Woodgate]

I read in sci.electronics.design that Walter Harley
<walterh@cafewalterNOSPAM.com> wrote (in
<4O-dnZ2dnZ3I8yLenZ2dne8gm96dnZ2dRVn-yJ2dnZ0@speakeasy.net&gt about
'transformer impedance', on Fri, 19 Aug 2005:

"Joerg" <notthisjoergsch@removethispacbell.net> wrote in message
newsJvNe.2687$AT7.444@newssvr22.news.prodigy.net...
That is what someone else in this forum also said (Martin?). But
transformers are still used when there is no other viable option. These
are mostly hardcore EMI and other interference situations.

Transformers are still widely used in pro audio - perhaps not in the
"pro-sumer" category of gear, but in quite a lot of higher-end gear; just
take a quick look through the ads in any pro audio magazine. This is partly
because of audio engineers' habit of running millivolt-level signals through
hundreds of feet of cable and still expecting S/N ratios of 80dB or better,
and partly because of the mystique of successful recordings made with
transformer gear in the past.

The problems with transformers, certainly, are that they are expensive and
large. But the common-mode rejection, ESD and EMI handling, distortion, and
noise specs of a good audio transformer are really quite hard to beat. (Bad
ones, on the other hand, are lousy.)


Is right, wot 'e say. I very rarely 'me, too', but this is exactly what

I would have written if WH hadn't.
--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

 

[Joerg]

Hello Martin,

I dont thik it was me. But generally I agree. There was a thread many
months ago about relays, me Vs Win, more or less. Cant remember who
"won" though. Probably Win

For audio a good transformer , say Jensen JE16 or Sowter equivalent,
in conjunction with star quad cable is incredible robust in terms of
CMRR. generally for mic levels. The bog standard single or triple
opamp circuitis usually adequate for line level signals.

This one is interesting for high CMRR if audio had to be transferred
across very long data lines like CAT-5, but seems not suited for very
low frequency:
http://www.muxlab.com/assets/files/datasheets/VE_AV_Analog.pdf

It got a friend out of a pickle once. They also make XLR pro stuff.

One test I have always meant to do is a CM 20KHz to100MHz AM, 50% mod.
sweep of opamps micamps (INA103 etc)to see how good they are under
difficult conditions, ie 60dB gain. This AFAIK is never published in
the spec sheets

You might be better off using differential video amps that have to
perform up to several MHz. My favorite was the uA733.

Regards, Joerg

http://www.analogconsultants.com

 

[John Woodgate]

I read in sci.electronics.design that Joerg
<notthisjoergsch@removethispacbell.net> wrote (in
<c3MNe.272$Ux3.242@newssvr21.news.prodigy.com&gt about 'transformer
impedance', on Sat, 20 Aug 2005:

This one is interesting for high CMRR if audio had to be transferred
across very long data lines like CAT-5, but seems not suited for very
low frequency:
http://www.muxlab.com/assets/files/datasheets/VE_AV_Analog.pdf

This item in the spec worries me:

Maximum Input Level 1.1Vp-p (+20 dBu, unbalanced < 1% THD)

It isn't just a typo. 1.1 Vp-p is -6 dBu.
--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

 

[Terry Given]

Ban wrote:

Tim Williams wrote:

The best of both worlds, particularly if you are transferring power,
is to match source to load. But you knew that! ;-)



Nope, this is not recommended. It will reduce the level by 6dB each stage
and waste a lot of heat unnecessarily. The same is true for audio
amplifiers. Their output impedance should approach 0, otherwise the
loudspeakers might develop additional errors in the frequency response.

Reductio ad absurdum: maximum power transfer applied to mains
distribution

Cheers
Terry

 

[Walter Harley]

"Tim Williams" <tmoranwms@charter.net> wrote in message
news:1ARMe.6900$rc6.6234@fe03.lga...

The best of both worlds, particularly if you are transferring power, is to
match source to load. But you knew that! ;-)

Here's my understanding:

Suppose you have a source with fixed output impedance Zs, and you want to
deliver as much power as possible to a load. What load impedance will work
best? Zero resistance won't work, because all the voltage will be dropped
across Zs, so delivered power will be P = EI = 0*I = 0. Infinite resistance
won't work, because there will be no current, so delivered power will be P =
E*0 = 0. If you do the math, you discover that the optimum load is the same
as Zs, at which point half the power is being dropped across the load and
half across the source.

On the other hand, suppose you have a *load* with fixed impedance Zl, and
you want to deliver as much power as possible to it. What source impedance
will work best? Zero resistance works great: the full voltage of the source
is delivered to the load, so power is P = EI = E^2/Zl. Any higher source
resistance just reduces the amount of current you can deliver, as well as
the voltage across the load. Note, though, that you want Zl to be
non-zero - otherwise it can't consume any power.

But that's all about optimizing power transfer. In audio-land, we could
care less about power transfer; what we want is accuracy, despite variations
in cable capacitance and load impedance, coupling of noise currents and
voltages, and so forth. For that goal too, zero source impedance and
moderate load impedance work well.

That's how I understand things - if I'm wrong, I'd be glad to be set right.

-walter

 

[John Woodgate]

I read in sci.electronics.design that Terry Given <my_name@ieee.org>
wrote (in <j68Oe.4444$iM2.462621@news.xtra.co.nz&gt about 'transformer
impedance', on Mon, 22 Aug 2005:

Reductio ad absurdum: maximum power transfer applied to mains
distribution

.... usually achieved with a pickaxe or digger bucket.
--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk

 

[John Woodgate]

I read in sci.electronics.design that Walter Harley
<walterh@cafewalterNOSPAM.com> wrote (in
<-rSdndoetcXC65TeRVn-qA@speakeasy.net&gt about 'power transfer (was:
transformer impedance)', on Sun, 21 Aug 2005:

That's how I understand things - if I'm wrong, I'd be glad to be set
right.

You are right.
--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk