[Iain Churches]

I am hoping to be able to borrow the necessary equipment
to make some phase measurements of various bits of tube
equipment. As far as RC coupled power amps are concerned,
I would expect the LF output to lead the input, and the
HF to lag it. Is this correct?

I cannot recall having seen any graphs from commercial
amp builders, but I am hoping that these figs, open and
closed loop might tell me something about amplifier
stability, which, for the homebrew builder of PP amps
with feedback is perhaps the greatest challenge of all:-)

Has anyone taken these kind of measurements? I would be
interested to hear what kind of figs have been plotted, and
what might be regarded as typical/good.


Festive greetings to all fellow rodents

Iain

[Patrick Turner]

Iain Churches wrote:

I am hoping to be able to borrow the necessary equipment
to make some phase measurements of various bits of tube
equipment. As far as RC coupled power amps are concerned,
I would expect the LF output to lead the input, and the
HF to lag it. Is this correct?


What does very basic LCR theory predict?
What were your observations with a dual trace CRO
with various R & C and R & L arrangements?
What is your knowledge of basic equivalent modelling
of amplifiers?

If you are searching for answers to the above, then you
now know how to spend your time off over christmas.
Out of the pub, and locked in your workshop I'd say.



I cannot recall having seen any graphs from commercial
amp builders, but I am hoping that these figs, open and
closed loop might tell me something about amplifier
stability, which, for the homebrew builder of PP amps
with feedback is perhaps the greatest challenge of all:-)

The phase response of many amplifiers both before and after NFB
was applied used to be published but not now because it
is of such little significance.



Has anyone taken these kind of measurements? I would be
interested to hear what kind of figs have been plotted, and
what might be regarded as typical/good.

Usually less than 20 degrees of phase shift maximum
anywhere across the band is regarded as OK.
Most amps with NFB easily achieve such criteria with a resistance load.

The R loaded amp phase response is not an indicator of
unconditional stability.

Just rememeber that the crossover filters used in most speakers
plus the drivers themselves add the majority of phase shift in a system.
There is much phase shifting in the recording process unless kept
all digital.

A study done in Scientic American in about 1967 concluded that
nobody could detect the relative phase of harmonics in a given
tone, so that if for example you have a 200Hz fundemental musical tone,
then if its 2H or 3H which could be 20% of the fundemental amplitude
was phase shifted relative to the fundemental by +/- 90 degrees,
nobody heard anything different; "the sound" didn't change.
The article concluded it was just as well our ears do not react
to phase too much, or else we would spend all day being
aurally muddled.
Evolution evolved ears which are very good at determining direction
of sound, since phase does matter, but audiophile concerns
about phase and its effect on "the sound" are largely twaddle.
Just don't let me catch you doing anything to wreck the phase though
when you record something.
Oh, and speakers often have the tweeters reverse phase connected, ie,
180 degrees
different to the midrange, and yet folks tell me that sounds better than
having everything "phase coherent".

By what amount do the realtive phase relationships of say 50 frequencies
present in an arriving musical signal at our ears change
when leaning forward say 400mm in a chair?

Patrick Turner.





Festive greetings to all fellow rodents

Iain

[[email protected]]

Iain Churches wrote:
I cannot recall having seen any graphs from commercial
amp builders, but I am hoping that these figs, open and
closed loop might tell me something about amplifier
stability, which, for the homebrew builder of PP amps
with feedback is perhaps the greatest challenge of all:-)

Has anyone taken these kind of measurements?

A guy named Nyquist published a paper on this about 75 years ago :-).
"Regeneration Theory", Bell System Technical Journal 11.1 (Jan 1932) p
126. Bode and Black were doing important stuff too. Obviously prior to
the Bode plot and the Nyquist diagram, negative feedback had been used
advantageously, but the simple diagrams turn all the complex stability
criteria into very beautiful pictures that just make sense.

Even though in today's classrooms Nyquist's work is talked about rather
abstractly in terms of information theory, this is a shame because he
was solving the most important on-the-ground important problems in
electronics and telegraphy and telephone technology. It certainly is
impressive that people still talk in fancy language about all the stuff
Nyquist was doing back then :-).

The Radiotron Designers Handbook 4th edition has a chapter nearly a
hundred pages long that walks you through every step for several
amplifier topologies, including (very importantly) measurements
necessary to draw the Nyquist diagram. Look in particular around page
338. Note when comparing old and modern references that the sign of the
imaginary axis will randomly flip. The long tabulations of complex math
from the middle of last century would probably be done on a spreadsheet
today :-).

The RDH is floating around the net, google for "RDH4", you want
specifically Chapter 7 but the rest of the book is wonderful too. I
have been lucky enough to locate several paper copies which I believe
is far more convenient than tying myself to a computer to read PDF's.

As to your original question of lead vs lag, for a typical RC coupled
amplifier yes there is a lead at LF and a lag at HF. Chapter 17 has
graphs for a "typical" Williamson. I put "typical" in quotes because I
do not believe that a real Williamson was really that flat, but then
again I never had his magical transformer.

Tim.

[John Byrns]

In article ,
Patrick Turner wrote:

Iain Churches wrote:

I am hoping to be able to borrow the necessary equipment
to make some phase measurements of various bits of tube
equipment. As far as RC coupled power amps are concerned,
I would expect the LF output to lead the input, and the
HF to lag it. Is this correct?

What does very basic LCR theory predict?
What were your observations with a dual trace CRO
with various R & C and R & L arrangements?
What is your knowledge of basic equivalent modelling
of amplifiers?

If you are searching for answers to the above, then you
now know how to spend your time off over christmas.
Out of the pub, and locked in your workshop I'd say.

I wonder what "equipment" Iain is hoping to borrow to make these
measurements? I always thought an audio oscillator and a dual channel
CRO were all you needed to make measurements more than good enough for
tube amplifier work? All one needs to do is simultaneously display the
input and output waveforms, read the relative time delay using the
graticule on the screen, and finally convert the relative time delay to
phase based on the frequency being measured.

A study done in Scientic American in about 1967 concluded that
nobody could detect the relative phase of harmonics in a given
tone, so that if for example you have a 200Hz fundemental musical tone,
then if its 2H or 3H which could be 20% of the fundemental amplitude
was phase shifted relative to the fundemental by +/- 90 degrees,
nobody heard anything different; "the sound" didn't change.
The article concluded it was just as well our ears do not react
to phase too much, or else we would spend all day being
aurally muddled.

You are in trouble now, audiophiles pride themselves on being able to
hear phase, and like to think their ears are special.

By what amount do the realtive phase relationships of say 50 frequencies
present in an arriving musical signal at our ears change
when leaning forward say 400mm in a chair?

Neglecting room effects, in other words if you are listening in an
anechoic chamber, leaning forward say 400mm in a chair has no effect on
the "relative phase relationships of say 50 frequencies". Leaning
forward only affects the time delay, and the time the CD spends siting
on your shelf has a much greater effect on the time delay.


Regards,

John Byrns

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

[Patrick Turner]

John Byrns wrote:

In article ,
Patrick Turner wrote:

Iain Churches wrote:

I am hoping to be able to borrow the necessary equipment
to make some phase measurements of various bits of tube
equipment. As far as RC coupled power amps are concerned,
I would expect the LF output to lead the input, and the
HF to lag it. Is this correct?

What does very basic LCR theory predict?
What were your observations with a dual trace CRO
with various R & C and R & L arrangements?
What is your knowledge of basic equivalent modelling
of amplifiers?

If you are searching for answers to the above, then you
now know how to spend your time off over christmas.
Out of the pub, and locked in your workshop I'd say.

I wonder what "equipment" Iain is hoping to borrow to make these
measurements? I always thought an audio oscillator and a dual channel
CRO were all you needed to make measurements more than good enough for
tube amplifier work? All one needs to do is simultaneously display the
input and output waveforms, read the relative time delay using the
graticule on the screen, and finally convert the relative time delay to
phase based on the frequency being measured.

A study done in Scientic American in about 1967 concluded that
nobody could detect the relative phase of harmonics in a given
tone, so that if for example you have a 200Hz fundemental musical tone,
then if its 2H or 3H which could be 20% of the fundemental amplitude
was phase shifted relative to the fundemental by +/- 90 degrees,
nobody heard anything different; "the sound" didn't change.
The article concluded it was just as well our ears do not react
to phase too much, or else we would spend all day being
aurally muddled.

You are in trouble now, audiophiles pride themselves on being able to
hear phase, and like to think their ears are special.

I have a flame suit on.

But many audiophiles have no better hearing than anyone average.
They have a very close mental connection to sound, perhaps they are
obsessive about it all....
Some are most definately more aurally perceptive than others, and their
brains make a lot more
from a given bit of music than average ppl, and some
understand each word of an opera where the avereage needs sub-titles,
and they hum along, since they have perfect pitch....
The significance of phase problems in amps may mot worry them too much.



By what amount do the realtive phase relationships of say 50 frequencies
present in an arriving musical signal at our ears change
when leaning forward say 400mm in a chair?

Neglecting room effects, in other words if you are listening in an
anechoic chamber, leaning forward say 400mm in a chair has no effect on
the "relative phase relationships of say 50 frequencies". Leaning
forward only affects the time delay, and the time the CD spends siting
on your shelf has a much greater effect on the time delay.

But the relative phase is the myriad of phase angles relative to a
reference
phase angle which could ultimately be related to the time of the day, 0
degrees
being midnight. The relative phase is all we are ever going to
experience meaning fully.
We listen to constantly changeing different relative phases of
frequencies all day long.
Even in an anechoic chamber the slightest head movement causes gross
relative phase
differences because every F has a different wave L.
We just notice about nothing, although we are gifted with mamal brains
which
have evolved to sense direction based on relative phase using TWO
microphones each side
of the head.
All the critters who didn't know where the lion was got eaten and their
attempts at
evolving hearing were curtailed.
Then the musicians move around a lot.......
It still sounds like an orchestra regardless of whether they sit still
as possible or
dance with their violins.
Without constant relative phase changes continuously, music fast becomes
boring,
like an organ without real vibrato, ie, shifting the note F
either side of the wanted note F.

So the more phase variation, the better.

Patrick Turner.




Regards,

John Byrns

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

[[email protected]]

Patrick Turner wrote:
The phase response of many amplifiers both before and after NFB
was applied used to be published but not now because it
is of such little significance.

Not much for listening, but if you are like the Iain and me and are
just learning the ropes of NFB and stability criteria I think there is
a lot to learn.

I learned a lot from your schematics on the web, for example. But I
learned much much more when I actually built some amps, tested and
measured and graphed phase and gain and Nyquist diagrams, and then
actually understood why you stuck those shelving networks in :-).

Tim.

[robert casey]

I wonder what "equipment" Iain is hoping to borrow to make these
measurements? I always thought an audio oscillator and a dual channel
CRO were all you needed to make measurements more than good enough for
tube amplifier work? All one needs to do is simultaneously display the
input and output waveforms, read the relative time delay using the
graticule on the screen, and finally convert the relative time delay to
phase based on the frequency being measured.


IN addition to the above, I'd take a look at a square wave. See how
well it gets reconstructed. If the various harmonics have the same time
delay (aka "group delay").

A study done in Scientic American in about 1967 concluded that
nobody could detect the relative phase of harmonics in a given
tone, so that if for example you have a 200Hz fundemental musical tone,
then if its 2H or 3H which could be 20% of the fundemental amplitude
was phase shifted relative to the fundemental by +/- 90 degrees,
nobody heard anything different; "the sound" didn't change.
The article concluded it was just as well our ears do not react
to phase too much, or else we would spend all day being
aurally muddled.

Yes for continious tones, but what about transient spikes? Music has
lots of those.

--
Posted via a free Usenet account from http://www.teranews.com

[Patrick Turner]

wrote:

Patrick Turner wrote:
The phase response of many amplifiers both before and after NFB
was applied used to be published but not now because it
is of such little significance.

Not much for listening, but if you are like the Iain and me and are
just learning the ropes of NFB and stability criteria I think there is
a lot to learn.

I learned a lot from your schematics on the web, for example. But I
learned much much more when I actually built some amps, tested and
measured and graphed phase and gain and Nyquist diagrams, and then
actually understood why you stuck those shelving networks in :-).

Tim.

Nearly all amps with any form of NFB need the gain and phase shift
reduced
just either side of the audio band of 20Hz to 20kHz.

But I re-itterate what my web-pages suggest and that is that
the actual compensation and shelving networks
need to be minimal in bandwidth limiting effect
and just the right values are chosen to get a good measured response
of about 20Hz to 35kHz, -1dB, R load, and still give
unconditional stabilty, ie, allow any kind of load
to be connected or no load at all, and without severe
peaking in the sine wave response within the AF band, or just outside
it,
as should be the case with a 2.0 uF load, which all amps should be able
to cope with at average power levels of 1/10 of clipping power levels.


Since one can't ever know EXACTLY what reactive components are with an
OPT
especially at HF, then it is pointless to spend days and days
on Nyquist charts for a one of a kind amp.
Just build the darn amp, and then cajole it to be stable
with educated guesses and careful trials and measurements
of LCR compo values until you have it right.
The shelving networks for HF are usually SO CRITICAL that
they are not called critical damping networks for nothing, so i don't
waste time calculating them; I'd only get the calculations wrong;
I just use what works with whatever
OPT and tube line up I have in front of me.
Even the poorest OPT can be used with triodes and allow 12dB of NFB
to tidy up behavioural loose ends
and to ensure the tubes produce music without grunge.

I hope my website examples continue to give a guide to all.

Patrick Turner.

[Patrick Turner]

Robert Casey wrote:


I wonder what "equipment" Iain is hoping to borrow to make these
measurements? I always thought an audio oscillator and a dual channel
CRO were all you needed to make measurements more than good enough for
tube amplifier work? All one needs to do is simultaneously display the
input and output waveforms, read the relative time delay using the
graticule on the screen, and finally convert the relative time delay to
phase based on the frequency being measured.


IN addition to the above, I'd take a look at a square wave. See how
well it gets reconstructed. If the various harmonics have the same time
delay (aka "group delay").

A study done in Scientic American in about 1967 concluded that
nobody could detect the relative phase of harmonics in a given
tone, so that if for example you have a 200Hz fundemental musical tone,
then if its 2H or 3H which could be 20% of the fundemental amplitude
was phase shifted relative to the fundemental by +/- 90 degrees,
nobody heard anything different; "the sound" didn't change.
The article concluded it was just as well our ears do not react
to phase too much, or else we would spend all day being
aurally muddled.

Yes for continious tones, but what about transient spikes? Music has
lots of those.

But there is nothing that happens faster in music than
a what happens in tone burst from nothing to full power of
a sine wave at 20kHz, mainly due to the
bandwidth limiting of the cd format, and certainly hi-fi
FM transmisions don't have anything above about 16kHz.

Try as hard as you like, but there will be no signal in any
audio amp which rises from zero volts to clipping voltage
in less than 1/100,000 of a second.
A 20kHz sine wave takes 1/20,000 seconds to happen.
The rise from zero V to the positive or negative peak
takes 1/80,000 seconds, so the slope at zero crossings is
such that zero volts to clipping should be doable in about
1/100,000 seconds, which = 10uS.

If an amp can make 32 watts into 8 ohms, that's 16Vrms,
or about 22.5 peak volts.
A rise of 22.5 peak volts in 10uS = 2.25V/uS


To ensure such performance can be attained, we try to build
our amps to be able to achieve a sine wave at 65kHz into the rated R
load
at up to 0.7 x 1 khz sine wave clipping voltage, and without having
any active device going into grid current/overloaded/saturated
condition.
I will leave you guys to figure the V/us the HF pole of 65kHz allows,
and to consider that an amp producing 128 watts into 8 ohms
needs to have a "rise time" = 2 x that of the amp making 32 watts.


If these conditions are met then the amp will withstand any "transient"
produced
in music with ease.
Transients in music are simply a complex array of HF sine waves
of various F and with various decay rates, but nothing
faster than can happen does happen.
Solid state amps ensure the output is bandwidth limited
due to the Zobel LR and CR networks, and these ensure the amp with a lot
of NFB
is never embarrassed by being forced to achieve absurdly
fast rise times which are simply not required.

Patrick Turner.




--
Posted via a free Usenet account from http://www.teranews.com

[Iain Churches]

I wonder what "equipment" Iain is hoping to borrow to make these
measurements? I always thought an audio oscillator and a dual channel
CRO were all you needed to make measurements more than good enough for
tube amplifier work? All one needs to do is simultaneously display the
input and output waveforms, read the relative time delay using the
graticule on the screen, and finally convert the relative time delay to
phase based on the frequency being measured.

Hi John,

Yes I have previously used the method you described. I want to plot
10Hz to 100kHz so the above method takes some time:-)

The system I am using is built by Feedback Electronics UK.
It consists of an oscillator with two outputs and two independent
gain controls, and a metering unit with two inputs.
(Yes, you have guessed it) You plug one of the oscillator
outputs straight into meter input A and the other into the amp. You
bring back the output from the amp to the meter input B which has
a digital read out to show you the phase difference. A LED shows
if the return signal is lagging or leading.

I have also used another method with the scope in XY mode, but
the dedicated 2-channel oscillator and meter is far quicker and
more accurate too.

Regards to all
Iain

[Iain Churches]

wrote in message
ups.com...
Patrick Turner wrote:
The phase response of many amplifiers both before and after NFB
was applied used to be published but not now because it
is of such little significance.

Not much for listening, but if you are like the Iain and me and are
just learning the ropes of NFB and stability criteria I think there is
a lot to learn.

Hi Tim,

Building a tube amp with quite an impressive performance is not
beyond the capability of most dedicated constructors, but learning
about NFB and critical damping is not so easy, but nevertheless
necessary if one wants to progress as a tube amp builder.

Iain

[Ian Iveson]

Tim wrote:

I learned a lot from your schematics on the web, for example. But I
learned much much more when I actually built some amps, tested and
measured and graphed phase and gain and Nyquist diagrams, and then
actually understood why you stuck those shelving networks in :-).


Did this give you any insight into how you might design a decent amp,
without shelving networks?

cheers, Ian

[Ian Iveson]

Tim wrote:

Even though in today's classrooms Nyquist's work is talked about
rather
abstractly in terms of information theory, this is a shame because
he
was solving the most important on-the-ground important problems in
electronics and telegraphy and telephone technology. It certainly is
impressive that people still talk in fancy language about all the
stuff
Nyquist was doing back then :-).

Not in my classrooms. Nyquist is general, foundation level
engineering.

As ever, I would recommend a decent student textbook on control system
theory. Such would place Nyquist and others (Bode, Laplace, Fourier,
etc) in a broad engineering context.

It is very useful to be able to see how all control systems have the
same problems, and can be designed with the same analytical tools.

cheers, Ian

[Patrick Turner]

Ian Iveson wrote:

Tim wrote:

I learned a lot from your schematics on the web, for example. But I
learned much much more when I actually built some amps, tested and
measured and graphed phase and gain and Nyquist diagrams, and then
actually understood why you stuck those shelving networks in :-).


Did this give you any insight into how you might design a decent amp,
without shelving networks?

cheers, Ian

Yes of course Ian, it did, and some amps with NFB need very little
shelving.

In amps with no global NFB can be the case with using triodes with a
high
OPT Z ratio to get a low Rout, one needs no shelving networks at all.

Also in SS amps with emitter or source follower outputs, there is no
need
for such networks when there is no NFB, for example see the amps built
by Sue Parker.

But nearly all SS amps have a C or R+C between the collector output and
base input
of the VAS to give a "dominant pole" for roll off of HF
at maybe only 500Hz and reducing at 6 dB/octave so that
for all of the AF band above about 5kHz the open loop phase shift has
a phase lag of -90 degrees.
The applied global NFB corrects virtually all of this delay.

In nearly all amplifiers unless there is a gain shelving network at HF
at least, then the NFB will become PFB at some F, and the poorer the
quality
of the OPT, then the lower the F at which this occurs, and it
is shown up by testing with pure capacitor loads which react with the
inductive
output resistance character of tube amps and their OPTs.

One may spend years modelling and drawing equivalent models
of amplifiers, or try to simulate them, but this is only
an exploration into principles; real amp makers know the principles and
concepts
and just apply the right values of R&C to eliminate any
chance of oscillation at any F under any load condition whatsoever.
Observation with a CRO and a volt meter leads them to the
one and only critical damping solution which is the best.
They will achieve this without allowing the amp to become
saturated and the HF performance is thus never sonically compromised.
Many haters of NFB don't fully understand the technical issues which
actually improve the sound rather than degrade it.

And even in UL amps or amps with local CFB in the OPT there will usually
be a need for critical damping to control HF behaviour.


But you would already know all about it I presume.


Patrick Turner.

[Iain Churches]

"John Byrns" wrote in message
...
I wonder what "equipment" Iain is hoping to borrow to make these
measurements? I always thought an audio oscillator and a dual channel
CRO were all you needed to make measurements more than good enough for
tube amplifier work? All one needs to do is simultaneously display the
input and output waveforms, read the relative time delay using the
graticule on the screen, and finally convert the relative time delay to
phase based on the frequency being measured.


Hi John,
Here is a pic of the Phase Oscillator made by
Feedback Electronics,.UK
http://www.kolumbus.fi/iain.churches...Oscillator.jpg

And here is a pic of the Phase Meter which goes with it:
http://www.kolumbus.fi/iain.churches...PhaseMeter.jpg

And here is a screen shot chart of the phase response I plotted for a
30W monobloc (Press F11 to see the graph full screen)
http://www.kolumbus.fi/iain.churches...seResponse.png

The system can be used in several ways, and one can take measurements
without the meter unit, but feeding the ref and return signals into the
X and Y inputs respectively of a CRO and then adjusting the phase of
the signal feeding the unit on test to give a straight line on the scope.
You can then read off the phase angle from the variable oscillator dial.

If you have both units, you can read the phase difference straight from the
meter.

Regards to all
Iain

[John Byrns]

In article ,
"Iain Churches" wrote:

"John Byrns" wrote in message
...
I wonder what "equipment" Iain is hoping to borrow to make these
measurements? I always thought an audio oscillator and a dual channel
CRO were all you needed to make measurements more than good enough for
tube amplifier work? All one needs to do is simultaneously display the
input and output waveforms, read the relative time delay using the
graticule on the screen, and finally convert the relative time delay to
phase based on the frequency being measured.


Hi John,
Here is a pic of the Phase Oscillator made by
Feedback Electronics,.UK
http://www.kolumbus.fi/iain.churches...Oscillator.jpg

Why is a special oscillator required, is it computer controlled or
something?

And here is a pic of the Phase Meter which goes with it:
http://www.kolumbus.fi/iain.churches...PhaseMeter.jpg

After you first mentioned the phase meter it occurred to me how simple
it would be to build a phase meter for sine waves. Simply feed the
reference signal and the signal to be measured through limiters. To
drive an analog meter just feed the outputs of two limiters through an
exclusive or logic gate, and use a D type flip-flop to drive a lead-lag
led. For a digital readout use a microprocessor with on chip
counter-timers to measure the period of the reference signal and the
time offset, compute the phase from that data and display it. Or if you
are more ambitious and like a challenge you could build some sort of
complex counter circuit from discrete SSI type IC's.

And here is a screen shot chart of the phase response I plotted for a
30W monobloc (Press F11 to see the graph full screen)
http://www.kolumbus.fi/iain.churches...seResponse.png

Was the data for this graph hand collected, or was an automated computer
program used? It would be useful if the graph also included the
amplitude response.

I'm not sure why I would want "to see the graph full screen", which is
probably a good thing because "F11" does something completely different
than show "the graph full screen".

The system can be used in several ways, and one can take measurements
without the meter unit, but feeding the ref and return signals into the
X and Y inputs respectively of a CRO and then adjusting the phase of
the signal feeding the unit on test to give a straight line on the scope.
You can then read off the phase angle from the variable oscillator dial.

If you have both units, you can read the phase difference straight from the
meter.


Regards,

John Byrns

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

[Iain Churches]

"John Byrns" wrote in message
...
In article ,
"Iain Churches" wrote:

"John Byrns" wrote in message
...
I wonder what "equipment" Iain is hoping to borrow to make these
measurements? I always thought an audio oscillator and a dual channel
CRO were all you needed to make measurements more than good enough for
tube amplifier work? All one needs to do is simultaneously display the
input and output waveforms, read the relative time delay using the
graticule on the screen, and finally convert the relative time delay to
phase based on the frequency being measured.


Here is a pic of the Phase Oscillator made by
Feedback Electronics,.UK
http://www.kolumbus.fi/iain.churches...Oscillator.jpg

Why is a special oscillator required, is it computer controlled or
something?

No. But it is intended as a stand-alone unit, and needs two outputs,
one of which needs to be variable phase if you measuring with the
scope and without the meter,

And here is a pic of the Phase Meter which goes with it:
http://www.kolumbus.fi/iain.churches...PhaseMeter.jpg

After you first mentioned the phase meter it occurred to me how simple
it would be to build a phase meter for sine waves. Simply feed the
reference signal and the signal to be measured through limiters. To
drive an analog meter just feed the outputs of two limiters through an
exclusive or logic gate, and use a D type flip-flop to drive a lead-lag
led. For a digital readout use a microprocessor with on chip
counter-timers to measure the period of the reference signal and the
time offset, compute the phase from that data and display it. Or if you
are more ambitious and like a challenge you could build some sort of
complex counter circuit from discrete SSI type IC's.

The earlier version of this system had an analogue meter. The one I
am using, with a digital display, is later.

And here is a screen shot chart of the phase response I plotted for a
30W monobloc (Press F11 to see the graph full screen)
http://www.kolumbus.fi/iain.churches...seResponse.png

Was the data for this graph hand collected, or was an automated computer
program used? It would be useful if the graph also included the
amplitude response.

I just noted the readings against frequency, and entered them into Excel
which plotted the chart. I do most of my charts that way.

I'm not sure why I would want "to see the graph full screen", which is
probably a good thing because "F11" does something completely different
than show "the graph full screen".

Hmm. On most computers .png screen captures seem to display at
reduced size, and some of the text may be illegible.
On my two PCs (both Dell) F11 gives a full screen pic.

Regards
Iain Churches