[Nelson Gietz]

Hi all,
I've been lurking here for several years without posting
mainly because my interest is in prewar radio restoration.
But Patrick's remarks about speakers raises a question.
He wrote:
"Indeed. Their acoustic result is not controlled with a NFB loop.

Every speaker is thus at the mercy of material behaviours.

And the result is that each design of speaker is equivalent to having a
multi channel graphic equaliser with random settings for along the band
with perfect transducer at the end.

Trying to work all the slides for a level response and avoid phase
problems is a major problem.

My question is...has anyone ever designed a speaker
with a smaller coil of a few turns on the voice coil, to
provide feedback related to the speaker's actual motion?
Seems it might be a way to smooth out speaker response.
Cheers,
Nelson

[Eeyore]

Nelson Gietz wrote:

My question is...has anyone ever designed a speaker
with a smaller coil of a few turns on the voice coil, to
provide feedback related to the speaker's actual motion?
Seems it might be a way to smooth out speaker response.

Yes, it's been done.

It's called motional feedback. Philips did it with a piezo sensor too IIRC. It's
never caught on so I expect there are too many problems with it.

Graham

[Arny Krueger]

"Eeyore" wrote in
message
Nelson Gietz wrote:

My question is...has anyone ever designed a speaker
with a smaller coil of a few turns on the voice coil, to
provide feedback related to the speaker's actual motion?
Seems it might be a way to smooth out speaker response.

Yes, it's been done.

Again, and again, and again. A quick google search for subwoofer feedback
turned up numerous examples, projects, patents, reviews, etc.

The first example I actually heard was done back in the early 1970s. It was
a EV 30W with acellerometer-based feedback.

It's called motional feedback. Philips did it with a
piezo sensor too IIRC.

A piezo sensor is probably an acellerometer. A sensing coil produces
velocity feedback. Either can be turned into positional feedback with
appropriate circuitry (one or two stages of electronic mathematical
integration).


It's never caught on so I expect
there are too many problems with it.

Velodyne has subwoofer products with feedback. I believe their first produce
went on the market in the early 1980s.

The big inherent problem is extending the feedback to frequencies above a
few 100 Hz. The sensing device tends to become decoupled from the cone as
the frequency rises.

[robert casey]

Nelson Gietz wrote:

Hi all,
I've been lurking here for several years without posting
mainly because my interest is in prewar radio restoration.
But Patrick's remarks about speakers raises a question.
He wrote:
"Indeed. Their acoustic result is not controlled with a NFB loop.

Speaker cones bouncing back and forth can produce some electrical energy
that the amp would see. A little like a dynamic microphone. And if the
amp has a NFB loop at the speaker terminals, some of that voice coil
induced signal would get into that NFB. Whether or not you want or can
make use of it, that's another question...

[Engineer]

On May 29, 3:31 pm, "Nelson Gietz" wrote:
Hi all,
I've been lurking here for several years without posting
mainly because my interest is in prewar radio restoration.
But Patrick's remarks about speakers raises a question.He wrote:

"Indeed. Their acoustic result is not controlled with a NFB loop.



Every speaker is thus at the mercy of material behaviours.

And the result is that each design of speaker is equivalent to having a
multi channel graphic equaliser with random settings for along the band
with perfect transducer at the end.

Trying to work all the slides for a level response and avoid phase
problems is a major problem.

My question is...has anyone ever designed a speaker
with a smaller coil of a few turns on the voice coil, to
provide feedback related to the speaker's actual motion?
Seems it might be a way to smooth out speaker response.
Cheers,
Nelson

AFAIK, it's been tried a number of times as motion f/b but I have no
tangible reports to hand.
Many years ago I tried it with a Stentorian 1012 speaker that had two
voice coils of 7.5 nominal ohms each - in series you get 15 ohms, in
parallel 3.75 ohms, IIRC. So, I ran one coil from an 8 ohms tube amp
(a SE 6V6, about 3 watts, all I could afford back then!) and used the
second voice coil for NFB. At that time I had no 'scope or sig. gen.
to measure the results and I can't remember what network I used in f/
b/ path - I think you get a velocity signal from the second VC so
you'd need to integrate to get position, nor do I recall how it
sounded when it was not oscillating. g Anyway, I quickly dropped
the idea in favour of simple voltage NFB.
Cheers,
Roger

[Patrick Turner]

Engineer wrote:

On May 29, 3:31 pm, "Nelson Gietz" wrote:
Hi all,
I've been lurking here for several years without posting
mainly because my interest is in prewar radio restoration.
But Patrick's remarks about speakers raises a question.He wrote:

"Indeed. Their acoustic result is not controlled with a NFB loop.



Every speaker is thus at the mercy of material behaviours.

And the result is that each design of speaker is equivalent to having a
multi channel graphic equaliser with random settings for along the band
with perfect transducer at the end.

Trying to work all the slides for a level response and avoid phase
problems is a major problem.

My question is...has anyone ever designed a speaker
with a smaller coil of a few turns on the voice coil, to
provide feedback related to the speaker's actual motion?
Seems it might be a way to smooth out speaker response.
Cheers,
Nelson

AFAIK, it's been tried a number of times as motion f/b but I have no
tangible reports to hand.
Many years ago I tried it with a Stentorian 1012 speaker that had two
voice coils of 7.5 nominal ohms each - in series you get 15 ohms, in
parallel 3.75 ohms, IIRC. So, I ran one coil from an 8 ohms tube amp
(a SE 6V6, about 3 watts, all I could afford back then!) and used the
second voice coil for NFB. At that time I had no 'scope or sig. gen.
to measure the results and I can't remember what network I used in f/
b/ path - I think you get a velocity signal from the second VC so
you'd need to integrate to get position, nor do I recall how it
sounded when it was not oscillating. g Anyway, I quickly dropped
the idea in favour of simple voltage NFB.
Cheers,
Roger

The idea of devoting a coil on a voicecoil former to provide NFB
or "motional feedback: a la Phillips isn't new but to do it on each
driver in a 3 way speaker system
and get the darn things to remain stable is rarely feasible, or
sensible,
and would cost makers a huge amount to implement while trying to compete
with great sounding systems
without the unreliable added complexity.

In fact when using nornmal FB arrangements with global NFB the speaker
voice coils ARE under considerable
control because any wayward motion or distortion generates a small
voltage appearing across the low output resistance of the
amplifier, and then appears at the NFB input port at the amp input, and
this signal
is amplified A times to oppose its own production.

Basically, the distortion currents generated in a voice coil are thus
shunted by the low Rout of the amp.

But the behaviour of the cone materials, ribbon, or ESL membrane are
beyound the grip of any
NFB, and a mic which picks up the sound and feeds it back is required.
Nice idea, but one cannot compensate for phase shifts and produce
stability.
It is difficult to implement in a 3 way speaker system.

I don't mind a world where NFB does not need to control each and every
single facet of audiological
endeavour.

Patrick Turner.

[Phread]

"robert casey" wrote in message hlink.net...
Nelson Gietz wrote:

Hi all,
I've been lurking here for several years without posting
mainly because my interest is in prewar radio restoration.
But Patrick's remarks about speakers raises a question.
He wrote:
"Indeed. Their acoustic result is not controlled with a NFB loop.

Speaker cones bouncing back and forth can produce some electrical energy that the amp would see. A little like a dynamic
microphone. And if the amp has a NFB loop at the speaker terminals, some of that voice coil induced signal would get into that
NFB. Whether or not you want or can make use of it, that's another question...

Kenwood had a series of amps back in the '70's or '80's with a 4 wire
speaker hookup - 2 heavy wires to drive the speaker & 2 lighter ones to
drive the amp's feedback circuitry. There was internal feedback in the
amp to control the gain and stabilize the amp, of course, but there was
also an independant feedback network connected to the 4 wire feedback
input. Hooking up the feedback circuit directly to the speaker did result in
better, tighter sound.

[Ian Iveson]

Arny Krueger wrote
... A piezo sensor is probably an acellerometer. A sensing coil
produces velocity feedback. Either can be turned into positional
feedback with appropriate circuitry (one or two stages of electronic
mathematical integration).
...

Why won't a simple inversion suffice?

cheers, Ian

[Eeyore]

Ian Iveson wrote:

Arny Krueger wrote
... A piezo sensor is probably an acellerometer. A sensing coil
produces velocity feedback. Either can be turned into positional
feedback with appropriate circuitry (one or two stages of electronic
mathematical integration).
...

Why won't a simple inversion suffice?

Oh Lord.

The Iveson IDIOT turns up like the proverbial bad penny to ask a STUPID
question.

Did you think they use *positive* feedback ?

Graham

[Arny Krueger]

"Ian Iveson" wrote in
message k

Arny Krueger wrote

... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can be
turned into positional feedback with appropriate
circuitry (one or two stages of electronic mathematical
integration). ...

Why won't a simple inversion suffice?

Never studied calculus or classical physics I take it.

This might get you started:

http://www.fearofphysics.com/Xva/xva.html

This is the more technical version:

http://www.ugrad.math.ubc.ca/coursed.../velocity.html

[Patrick Turner]

Eeyore wrote:

Ian Iveson wrote:

Arny Krueger wrote
... A piezo sensor is probably an acellerometer. A sensing coil
produces velocity feedback. Either can be turned into positional
feedback with appropriate circuitry (one or two stages of electronic
mathematical integration).
...

Why won't a simple inversion suffice?

Oh Lord.

The Iveson IDIOT turns up like the proverbial bad penny to ask a STUPID
question.

Did you think they use *positive* feedback ?

But once upon a time Bogen amps had variable speaker damping.

How? posititive current FB adjustment.

They said it was good to control reactive bass speakers on transients

Its never used now, despite being easy to use,
but positive current FB lowers amplifier resistance
and can make any amp have Rout = 0.0 ohms, or even some negative number
of ohms,
ie, the output voltage rises when the load ohms are reduced; queer.
This can be dangerous because if the load is a short circuit such amps
can oscillate to death easily, with too much PCFB.

PCFB reduces BW and increases THD/IMD, but reduces Rout, unlike
positive voltage FB which also reduces BW and increases THD/IMD,
and increases Rout. But PVFB in the presence of negative current FB
and sprinkeled with a pinch of PCFB, sauteed with dash of local NCFB can
make a working amp,
until some idiot does a little mod somewhere, and then all hell breaks
loose.

Only NVFB and local NCFB is used these days.

Patrick Turner.



Graham

[Ian Iveson]

Arny Krueger wrote

... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can be
turned into positional feedback with appropriate
circuitry (one or two stages of electronic mathematical
integration). ...

Why won't a simple inversion suffice?

Never studied calculus or classical physics I take it.

This might get you started:

http://www.fearofphysics.com/Xva/xva.html

This is the more technical version:

http://www.ugrad.math.ubc.ca/coursed.../velocity.html

What's classical physics?

I don't think much of those pages. I suspect there is a key difference
between exhibiting knowledge, and teaching. Do you have a link that
starts off by explaining what calculus is, written by someone who can
teach by writing?

Anyway, I asked because they said at school that if I integrated a
sine I would get a cosine. Then they said that if I integrated the
cosine I would get an inverted sine. Years later, someone told me that
music is made of a sum of sines, and I took the liberty of thinking
that a series of sums could be integrated term by term.

So what's the problem with that?

cheers, Ian

[Ian Iveson]

Eeyore wrote

... A piezo sensor is probably an acellerometer. A sensing coil
produces velocity feedback. Either can be turned into positional
feedback with appropriate circuitry (one or two stages of
electronic
mathematical integration).
...

Why won't a simple inversion suffice?

Oh Lord.

The Iveson IDIOT turns up like the proverbial bad penny to ask a
STUPID
question.

Did you think they use *positive* feedback ?

Graham

You should learn to think before blurting. Perhaps you did. Sigh...

The devil is in the amplitude, not the phase.

Ian

[Eeyore]

Ian Iveson wrote:

What's classical physics?

Clearly something that passed you by.

Newton's Laws, Faraday, that kind of thing.

Graham

[Eeyore]

Ian Iveson wrote:

The devil is in the amplitude, not the phase.

Wrong AGAIN !

Graham

[Ian Iveson]

Eeyore wrote

What's classical physics?

Clearly something that passed you by.

Newton's Laws, Faraday, that kind of thing.

I see. So it's classical as in music, rather than in poetry or
philosophy or mathematics or those other Greek and Arab and Roman
things?

Maybe it's something to do with using instruments.

When did the classical period of physics begin and end? What followed,
and what went before? Was there an epoch of romance, or is that yet to
come?

cheers, Ian

[Eeyore]

Ian Iveson wrote:

Eeyore wrote

What's classical physics?

Clearly something that passed you by.

Newton's Laws, Faraday, that kind of thing.

I see. So it's classical as in music, rather than in poetry or
philosophy or mathematics or those other Greek and Arab and Roman
things?

Maybe it's something to do with using instruments.

When did the classical period of physics begin and end? What followed,
and what went before? Was there an epoch of romance, or is that yet to
come?

Classical physics if you really need it explaining is the stuff that deals
mainly with solid objects, things you can touch, the forces on things etc....

As opposed to quantum mechanics, particle physics and the like.

Graham

[Arny Krueger]

"Ian Iveson" wrote in
message k
Arny Krueger wrote

... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can be
turned into positional feedback with appropriate
circuitry (one or two stages of electronic mathematical
integration). ...

Why won't a simple inversion suffice?

Never studied calculus or classical physics I take it.

This might get you started:

http://www.fearofphysics.com/Xva/xva.html

This is the more technical version:

http://www.ugrad.math.ubc.ca/coursed.../velocity.html

What's classical physics?

Something that would really help you Ian, if you could even grasp the need
and application.

It explains why electrical integration is the right answer to the problem at
hand, and why simple inversion isn't.

I don't think much of those pages.

The real world is not a popularity contest.

If you don't like the alternatives I picked, then google up something that
fits you more.

I suspect there is a
key difference between exhibiting knowledge, and
teaching.

Not just a suspicion, the difference between exhibiting knowlege and
teaching is a widely-recognized fact.

However, some people are unteachable.

Do you have a link that starts off by
explaining what calculus is, written by someone who can
teach by writing?

First show me someone who can learn by reading.

Anyway, I asked because they said at school that if I
integrated a sine I would get a cosine. Then they said
that if I integrated the cosine I would get an inverted
sine. Years later, someone told me that music is made of
a sum of sines, and I took the liberty of thinking that a
series of sums could be integrated term by term.

So what's the problem with that?

Thanks Ian for the quick example of difference between exhibiting knowlege
which is what you just did, and knowing how to apply it in the real world
which is what is needed here.

[Ian Iveson]

Arny Krueger wrote

... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can be
turned into positional feedback with appropriate
circuitry (one or two stages of electronic mathematical
integration). ...

Why won't a simple inversion suffice?

Never studied calculus or classical physics I take it.

This might get you started:

http://www.fearofphysics.com/Xva/xva.html

This is the more technical version:

http://www.ugrad.math.ubc.ca/coursed.../velocity.html

What's classical physics?

Something that would really help you Ian, if you could even grasp
the need and application.

It explains why electrical integration is the right answer to the
problem at hand, and why simple inversion isn't.

I don't think much of those pages.

The real world is not a popularity contest.

If you don't like the alternatives I picked, then google up
something that fits you more.

I suspect there is a
key difference between exhibiting knowledge, and
teaching.

Not just a suspicion, the difference between exhibiting knowlege and
teaching is a widely-recognized fact.

However, some people are unteachable.

Do you have a link that starts off by
explaining what calculus is, written by someone who can
teach by writing?

First show me someone who can learn by reading.

Anyway, I asked because they said at school that if I
integrated a sine I would get a cosine. Then they said
that if I integrated the cosine I would get an inverted
sine. Years later, someone told me that music is made of
a sum of sines, and I took the liberty of thinking that a
series of sums could be integrated term by term.

So what's the problem with that?

Thanks Ian for the quick example of difference between exhibiting
knowlege which is what you just did, and knowing how to apply it in
the real world which is what is needed here.

I take it you don't know the answer to my question. Thanks anyway.

cheers, Ian

[Arny Krueger]

"Ian Iveson" wrote in
message k
Arny Krueger wrote

... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can
be turned into positional feedback with appropriate
circuitry (one or two stages of electronic
mathematical integration). ...

Why won't a simple inversion suffice?

Never studied calculus or classical physics I take it.

This might get you started:

http://www.fearofphysics.com/Xva/xva.html

This is the more technical version:

http://www.ugrad.math.ubc.ca/coursed.../velocity.html

What's classical physics?

Something that would really help you Ian, if you could
even grasp the need and application.

It explains why electrical integration is the right
answer to the problem at hand, and why simple inversion
isn't.
I don't think much of those pages.

The real world is not a popularity contest.

If you don't like the alternatives I picked, then google
up something that fits you more.

I suspect there is a
key difference between exhibiting knowledge, and
teaching.

Not just a suspicion, the difference between exhibiting
knowlege and teaching is a widely-recognized fact.

However, some people are unteachable.

Do you have a link that starts off by
explaining what calculus is, written by someone who can
teach by writing?

First show me someone who can learn by reading.

Anyway, I asked because they said at school that if I
integrated a sine I would get a cosine. Then they said
that if I integrated the cosine I would get an inverted
sine. Years later, someone told me that music is made of
a sum of sines, and I took the liberty of thinking that
a series of sums could be integrated term by term.

So what's the problem with that?

Thanks Ian for the quick example of difference between
exhibiting knowlege which is what you just did, and
knowing how to apply it in the real world which is what
is needed here.

I take it you don't know the answer to my question.
Thanks anyway.

Typical of your attitude problem, Ian. Of course I know the answer, but it
is not worth my trouble to try to hand-feed it to you, as it us is a bit of
work to explain this to someone with no serious interest in, or practical
knowlege of classical physics and calculus.

[Eeyore]

Arny Krueger wrote:

"Ian Iveson" wrote

I take it you don't know the answer to my question.
Thanks anyway.

Typical of your attitude problem, Ian. Of course I know the answer, but it
is not worth my trouble to try to hand-feed it to you, as it us is a bit of
work to explain this to someone with no serious interest in, or practical
knowlege of classical physics and calculus.

I can't help but wonder if he now wants to know what 'calculus' is. ;~)

Graham

[Ian Iveson]

Arny Krueger wrote

... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can
be turned into positional feedback with appropriate
circuitry (one or two stages of electronic
mathematical integration). ...

Why won't a simple inversion suffice?

Never studied calculus or classical physics I take it.

This might get you started:

http://www.fearofphysics.com/Xva/xva.html

This is the more technical version:

http://www.ugrad.math.ubc.ca/coursed.../velocity.html

What's classical physics?

Something that would really help you Ian, if you could
even grasp the need and application.

It explains why electrical integration is the right
answer to the problem at hand, and why simple inversion
isn't.
I don't think much of those pages.

The real world is not a popularity contest.

If you don't like the alternatives I picked, then google
up something that fits you more.

I suspect there is a
key difference between exhibiting knowledge, and
teaching.

Not just a suspicion, the difference between exhibiting
knowlege and teaching is a widely-recognized fact.

However, some people are unteachable.

Do you have a link that starts off by
explaining what calculus is, written by someone who can
teach by writing?

First show me someone who can learn by reading.

Anyway, I asked because they said at school that if I
integrated a sine I would get a cosine. Then they said
that if I integrated the cosine I would get an inverted
sine. Years later, someone told me that music is made of
a sum of sines, and I took the liberty of thinking that
a series of sums could be integrated term by term.

So what's the problem with that?

Thanks Ian for the quick example of difference between
exhibiting knowlege which is what you just did, and
knowing how to apply it in the real world which is what
is needed here.

I take it you don't know the answer to my question.
Thanks anyway.

Typical of your attitude problem, Ian. Of course I know the answer,
but it is not worth my trouble to try to hand-feed it to you, as it
us is a bit of work to explain this to someone with no serious
interest in, or practical knowlege of classical physics and
calculus.

Surely the answer could be simply and clearly stated in a line or two?

And what, incidentally, is "electronic mathematical integration"?

I can't see where I have provoked such a defensive response.

All good fortune,

Ian

[Chris Hornbeck]

On Wed, 30 May 2007 23:36:40 GMT, "Ian Iveson"
wrote:

I can't see where I have provoked such a defensive response.

The lack of any real response to your question is notable.
Part of the problematical lack of response is that

"... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can be
turned into positional feedback with appropriate
circuitry (one or two stages of electronic mathematical
integration). ..."

is literally incorrect. No translation is possible.
Acceleration cannot be mapped into position, and it's
not done that way.

Your questions are always interesting; don't stop.
Much thanks, as always,

Chris Hornbeck

[Arny Krueger]

"Ian Iveson" wrote in
message k
Arny Krueger wrote

... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can
be turned into positional feedback with appropriate
circuitry (one or two stages of electronic
mathematical integration). ...

Why won't a simple inversion suffice?

Never studied calculus or classical physics I take
it. This might get you started:

http://www.fearofphysics.com/Xva/xva.html

This is the more technical version:

http://www.ugrad.math.ubc.ca/coursed.../velocity.html

What's classical physics?

Something that would really help you Ian, if you could
even grasp the need and application.

It explains why electrical integration is the right
answer to the problem at hand, and why simple inversion
isn't.
I don't think much of those pages.

The real world is not a popularity contest.

If you don't like the alternatives I picked, then
google up something that fits you more.

I suspect there is a
key difference between exhibiting knowledge, and
teaching.

Not just a suspicion, the difference between exhibiting
knowlege and teaching is a widely-recognized fact.

However, some people are unteachable.

Do you have a link that starts off by
explaining what calculus is, written by someone who
can teach by writing?

First show me someone who can learn by reading.

Anyway, I asked because they said at school that if I
integrated a sine I would get a cosine. Then they said
that if I integrated the cosine I would get an
inverted sine. Years later, someone told me that
music is made of a sum of sines, and I took the
liberty of thinking that a series of sums could be
integrated term by term.

So what's the problem with that?

Thanks Ian for the quick example of difference between
exhibiting knowlege which is what you just did, and
knowing how to apply it in the real world which is what
is needed here.

I take it you don't know the answer to my question.
Thanks anyway.

Typical of your attitude problem, Ian. Of course I know
the answer, but it is not worth my trouble to try to
hand-feed it to you, as it is a bit of work to
explain this to someone with no serious interest in, or
practical knowlege of classical physics and calculus.

Surely the answer could be simply and clearly stated in a
line or two?

Some things aren't that easy. Some of the best minds in the business have
studied "mathematical integration", and it took them a lot more than two
lines of text to explain it. See below. Besides, given what we know about
you now, their work is probably inadequate to overcome your mental
situation.

And what, incidentally, is "electronic mathematical
integration"?

Try searching google for "mathematical integration". The Wikipedia article
that comes up very early in the list looks pretty good. But, it is a lot
more than a "line or two".

I'll take a big gamble here Ian, and hope that despite your inability to do
simple textual research on the web, you can figure out how the word
"electronic" fits into the phrase "electronic mathematical integration"?

I can't see where I have provoked such a defensive
response.

It's not defensiveness Ian, its accuracy and impatience with mental
laziness. Of course you don't see your problem. Please start out by
reviewing your snotty insult about me not knowing the answer to your
question.

[Arny Krueger]

"Chris Hornbeck" wrote in
message
On Wed, 30 May 2007 23:36:40 GMT, "Ian Iveson"
wrote:

I can't see where I have provoked such a defensive
response.

The lack of any real response to your question is notable.
Part of the problematical lack of response is that

"... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can be
turned into positional feedback with appropriate
circuitry (one or two stages of electronic mathematical
integration). ..."

is literally incorrect.

??????????????

No translation is possible.

??????????????

Acceleration cannot be mapped into position,

??????????????

and it's not done that way.

??????????????

I'm looking forward to answers to the above questions.

Here are some standard answers from a variety of independent sources:

(1) In fact accelerometers often use piezo elements.

http://www.dliengineering.com/vibman...elerometer.htm

"The piezo-electric accelerometer can be considered the standard vibration
transducer for machine vibration measurement."

(2) Mathematically integrating acceleration twice with suitable initial
conditions is a standard means for calculating position.

http://www.ugrad.math.ubc.ca/coursed.../velocity.html

(3) Mathematically integrating acceleration twice with suitable initial
conditions is a commonly-used means for calculating positional feedback for
loudspeakers.

http://www.analog.com/en/content/0,2...5F7004,00.html

"Our initial test setup consisted of a small dominant-pole amplifier
powering a 20 watt woofer fitted with a 5g accelerometer. The summing
junction was implemented by feeding the input signal into the non-inverting
input of an op-amp, and running the feedback to the inverting input.
Integration was accomplished with two simple op-amp integrators (Figure 1):"

While researching the third answer I found that a number of refinements have
been implemented. One of the more elegant ones enhanced the feedback
calculations with an adaptive filter that was implemented in the digital
domain.

[Ian Iveson]

Chris Hornbeck wrote

I can't see where I have provoked such a defensive response.

The lack of any real response to your question is notable.
Part of the problematical lack of response is that

"... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can be
turned into positional feedback with appropriate
circuitry (one or two stages of electronic mathematical
integration). ..."

is literally incorrect. No translation is possible.
Acceleration cannot be mapped into position, and it's
not done that way.

Your questions are always interesting; don't stop.
Much thanks, as always,

I hold you entirely responsible. Your efforts to present a no-nonsense
approach to stability in feedback amplifiers had me wondering how
likely it might be that a reader would know the nature of what we call
delay, and yet not in consequence appreciate the meaning of Bode. Slim
chance, it seemed to me at the time.

So I was trying to think of a simple experiment that everyone could do
at home. I considered employing a pendulum, because it's not
electrical and many older folk like us will have done the maths at
school, before calculus was abandoned in favour of numerical methods.
I reckoned I might progress from the idea of an equation to a transfer
function by wiggling the fulcrum at various frequencies by hand, so to
get a feel for how the pendulum responds in the shorter and longer
terms, thus distinguishing along the way between the transient and the
steady state response. Then a damped pendulum, using a crumpled tissue
for a bob, then a string of damped pendulums with various or similar
lengths, and finally posing the question of how easy it might be, if
you concentrated and tried to apply feedback to correct error, to sign
your name if you mounted a pen on the bottom bob. I was stuck,
however, on how to weave zeros, as well as poles, into the plot
without confusing myself.

Consequently the simple equation for harmonic motion, which states
that deceleration is proportional to distance, was fresh in my mind.

I don't think Arny has read the links he posted, BTW. Those students
struggling with the accelerometer aren't learning much. Sometimes a
real experiment is unwise. Their mentor should have refused to say
another word until they wrote something that made sense about where
they were. Then they might have realised that where they got to in the
end is still wrong.

thanks, Ian

[Chris Hornbeck]

On Thu, 31 May 2007 07:41:00 -0400, "Arny Krueger"
wrote:

"... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can be
turned into positional feedback with appropriate
circuitry (one or two stages of electronic mathematical
integration). ..."

is literally incorrect.

??????????????

No translation is possible.

??????????????

Acceleration cannot be mapped into position,

??????????????

and it's not done that way.

??????????????

I'm looking forward to answers to the above questions.

If you'd accept as given that I'm perfectly capable of
integrating dV/dT over time and am not unaware of the theory,
you might move on to examining what I've already said.

It was not posted trivially or argumentatively, but rather
as a suggested path to a useful model. Simple series'd
integrations are not a useful model because they don't
work right. Hint: they lack a large somewhat non-linear
constant *and* a zero-crossing flat spot.

Much thanks, as always,

Chris Hornbeck

[Arny Krueger]

"Chris Hornbeck" wrote in
message
On Thu, 31 May 2007 07:41:00 -0400, "Arny Krueger"
wrote:

"... A piezo sensor is probably an acellerometer. A
sensing coil produces velocity feedback. Either can
be turned into positional feedback with appropriate
circuitry (one or two stages of electronic
mathematical integration). ..."

is literally incorrect.

??????????????

No translation is possible.

??????????????

Acceleration cannot be mapped into position,

??????????????

and it's not done that way.

??????????????

I'm looking forward to answers to the above questions.

If you'd accept as given that I'm perfectly capable of
integrating dV/dT over time and am not unaware of the
theory,

Never said otherwise.

you might move on to examining what I've already said.

Which seems to lack substance.

It was not posted trivially or argumentatively,

If you say so, but you forgot the part about actually saying something
helpful.

but rather as a suggested path to a useful model.

Now I'm beginning to think that I missed a post.


Simple series'd
integrations are not a useful model because they don't
work right.

Depends on how you define *right*. Right now the leading proposed
alternative is simple inversion.

Hint: they lack a large somewhat non-linear
constant *and* a zero-crossing flat spot.

Sorry Chris but if you want to puff yourself up and play guru, its not my
day for playing the child.

Much thanks, as always,

Simple hypocracy by means of a canned sig line, in this case.

[Arny Krueger]

"Ian Iveson" wrote in
message k

I don't think Arny has read the links he posted, BTW.

As usual Ian, you're wrong.

Those students struggling with the accelerometer aren't
learning much.

All the people who made effective feedback subwoofers that way might be
laughing. A few have laughed on the way to the bank.

Sometimes a real experiment is unwise.

Sometimes a little practical experience can be a big help. Tell us about
yours, Ian.

[Ian Iveson]

Arny said:

Right now the leading proposed alternative is simple inversion.

Who proposed that? And what would be wrong with it?

Ian

[Chris Hornbeck]

On Fri, 1 Jun 2007 08:13:05 -0400, "Arny Krueger"
wrote:

Simple series'd
integrations are not a useful model because they don't
work right.

Depends on how you define *right*. Right now the leading proposed
alternative is simple inversion.

Proposed? Not by me. In fact, a working model is pretty
difficult. Series'd integrations are the basis of a large-
signal static model, but don't provide any sort of a real
general model.

Hint: they lack a large somewhat non-linear
constant *and* a zero-crossing flat spot.

Sorry Chris but if you want to puff yourself up and play guru, its not my
day for playing the child.

The hint is for anyone interested in exploring the topic.
No personal offense was intended, and in fact, I still
can't see it. I am, of course, sorry if offense was taken.

Much thanks, as always,

Simple hypocracy by means of a canned sig line, in this case.

Not canned; hand typed, new each post. Just for you:

Very much thanks, as always,

Chris Hornbeck

[Ian Iveson]

Arny Krueger wrote

I don't think Arny has read the links he posted, BTW.

As usual Ian, you're wrong.

Are you sure you read it properly, all the way to the end?

Those students struggling with the accelerometer aren't
learning much.

All the people who made effective feedback subwoofers that way might
be laughing. A few have laughed on the way to the bank.


That way? What way?

Sometimes a real experiment is unwise.

Sometimes a little practical experience can be a big help.

And sometimes not. We all experience life. To paraphrase Lenin,
experience is of limited value in the absence of coherent theory.

Tell us about yours, Ian.

With accelerometers in control systems applications? No. I don't do
personal life history stuff here.

With accelerometers in audio speakers? None, that's a reproductionist
engineer's dream, not mine. I build for music. In any case, there are
better ways of dealing with the foibles of speakers, except possibly
for low bass. Speakers behave predictably, after all. Dead reckoning
should be fine.

Ian

[Andre Jute]

In an otherwise dull technical thread, already spiralling into ennui,
Ian Iveson wrote:
To paraphrase Lenin,
experience is of limited value in the absence of coherent theory.

Ian

That is why Lenin screwed up an entire country. Russia under the last
Tsar (or more particularly his prime minister) was the fastest
industrializing country in Europe. Lenin had lots of theories but no
useful ideas. He would jail or execute his political opponents, then
appropriate their policies as his own. For instance the so-called New
Economic Policy, a reversion to a form of agri-capitalism, installed
by Lenin and Stalin after their original policies brought only chaos
(1), originated not with the Bolsheviks but with the Social
Revolutionaries.

It is clear to anyone with half a brain that a theory isn't worth ****
in the absence of experience. The purpose of experience is to provide
material for observation and the deduction or reduction of a theory to
provide a framework for the experience, and predict and improve future
experience. Lenin, as usual, perversely put the cart before the horse.
What a ****wit.

By the way, almost the last coherent words Lenin uttered before he
descended into his final syphilitic dementia were, "There aren't
enough executions." Typical.

We could do with a better class of thinking than Lenin's on RAT.

Andre Jute
No real corpses were harmed in the assembly of my golem Worthless
Wieckless -- CE Statement of Conformity

(1) By far the most amusing way to study the Bolshevik ****ups is as
economic history rather than as criminal history. You need a taste for
the bizarre, and to recognize surrealism wherever you find it, or you
won't believe how incompetent they were. To run a large,
industrializing nation's economy, for instance, Lenin telegraphed his
agent in Stockholm to send him half a dozen accountants; the agent was
told he could "fix their remuneration in accordance with local
conditions." To finance the treasury, Lenin sent thugs to the state
issuing bank to confiscate all the money they could find; Lenin
himself sat up part of the night in the Smolny (a girl's schools with
the Imperial crest still over the gate), waving a revolver over the
bag of money; presumably he didn't trust his associates! The communist
bosses were the Keystone Kops of economics; they'd be a riotous comedy
act if they didn't kill many more by starvation than they did by the
bullet in the back of the neck.

[Patrick Turner]

Andre Jute wrote:

In an otherwise dull technical thread, already spiralling into ennui,
Ian Iveson wrote:
To paraphrase Lenin,
experience is of limited value in the absence of coherent theory.

Ian

That is why Lenin screwed up an entire country. Russia under the last
Tsar (or more particularly his prime minister) was the fastest
industrializing country in Europe. Lenin had lots of theories but no
useful ideas. He would jail or execute his political opponents, then
appropriate their policies as his own. For instance the so-called New
Economic Policy, a reversion to a form of agri-capitalism, installed
by Lenin and Stalin after their original policies brought only chaos
(1), originated not with the Bolsheviks but with the Social
Revolutionaries.

It is clear to anyone with half a brain that a theory isn't worth ****
in the absence of experience. The purpose of experience is to provide
material for observation and the deduction or reduction of a theory to
provide a framework for the experience, and predict and improve future
experience. Lenin, as usual, perversely put the cart before the horse.
What a ****wit.

By the way, almost the last coherent words Lenin uttered before he
descended into his final syphilitic dementia were, "There aren't
enough executions." Typical.

We could do with a better class of thinking than Lenin's on RAT.

Andre Jute
No real corpses were harmed in the assembly of my golem Worthless
Wieckless -- CE Statement of Conformity

(1) By far the most amusing way to study the Bolshevik ****ups is as
economic history rather than as criminal history. You need a taste for
the bizarre, and to recognize surrealism wherever you find it, or you
won't believe how incompetent they were. To run a large,
industrializing nation's economy, for instance, Lenin telegraphed his
agent in Stockholm to send him half a dozen accountants; the agent was
told he could "fix their remuneration in accordance with local
conditions." To finance the treasury, Lenin sent thugs to the state
issuing bank to confiscate all the money they could find; Lenin
himself sat up part of the night in the Smolny (a girl's schools with
the Imperial crest still over the gate), waving a revolver over the
bag of money; presumably he didn't trust his associates! The communist
bosses were the Keystone Kops of economics; they'd be a riotous comedy
act if they didn't kill many more by starvation than they did by the
bullet in the back of the neck.

Lenin made a few speaches, and so he was a speaker.

History gave him all the feedback he was due to get....

Pools of blood, ruined dreams and lives......

Now the capitalist speakers are in full roar,
and the prosperity will cause planetary feedback to occur
with rising temperatures.....

The condition of mankind is that it never knows really where its at.

Womenkind are no help to the problem.

Patrick Turner.

[Arny Krueger]

"Ian Iveson" wrote in
message
Arny said:

Right now the leading proposed alternative is simple
inversion.

Who proposed that?

Iain Iveson.

And what would be wrong with it?

Utterly incapable of developing positional feedback from an accelerometer.

[Arny Krueger]

"Ian Iveson" wrote in
message k
Arny Krueger wrote

I don't think Arny has read the links he posted, BTW.

As usual Ian, you're wrong.

Are you sure you read it properly, all the way to the end?

Of course. Can you comprehend it?

Those students struggling with the accelerometer aren't
learning much.

All the people who made effective feedback subwoofers
that way might be laughing. A few have laughed on the
way to the bank.

That way? What way?

multiple integration.

Sometimes a real experiment is unwise.

Sometimes a little practical experience can be a big
help.

And sometimes not. We all experience life. To paraphrase
Lenin, experience is of limited value in the absence of
coherent theory.

Tell us about yours, Ian.

With accelerometers in control systems applications? No.
I don't do personal life history stuff here.

IOW, you have nothing to bring to the table.

With accelerometers in audio speakers? None, that's a
reproductionist engineer's dream, not mine.

Woooo! A gratuitous use of a 50 cent word.

I build for
music. In any case, there are better ways of dealing with
the foibles of speakers, except possibly for low bass.


Ironically, low bass is the primary application for servo-feedback woofers.

Speakers behave predictably, after all.

Yes, in dream land speakers are predicatable and consistent.

Dead reckoning should be fine.

Right, actually looking out the window to see where you are is for people
who don't properly respect deep thought and theories.

[Eeyore]

Arny Krueger wrote:

"Ian Iveson" wrote

And what would be wrong with it?

Utterly incapable of developing positional feedback from an accelerometer.

Impossible in any AC coupled network of course.

Something like a laser interferometer to measure motor displacement would be
interesting.

Graham

[Ian Iveson]

Arny said:

Right now the leading proposed alternative is simple
inversion.

Who proposed that?

Iain Iveson.

Who? Me, Ian Iveson?

I asked this question:

Why won't a simple inversion suffice?

....that so far no-one has answered.

Then I asked this question:

Anyway, I asked because they said at school that if I
integrated a sine I would get a cosine. Then they said
that if I integrated the cosine I would get an inverted
sine. Years later, someone told me that music is made of
a sum of sines, and I took the liberty of thinking that a
series of sums could be integrated term by term.

So what's the problem with that?

....that so far no-one has answered. Except now you have said:

Utterly incapable of developing positional feedback from an
accelerometer.

Which is silly.

I am deeply disappointed with you now, Arny.

There is an immediately obvious problem that you could have explained
in a few words. What interests me however is another set of
difficulties which arise from the nature of music. It could be a good
discussion in the right place. Particularly when Patrick's point about
current feedback is added to the mix. Sadly this is no longer the
right place.

Ian

PS Actually they told me that if I integrate a sine I will get an
inverted cosine, and if I integrate that I'll get an inverted sine.
Incidentally, if I integrate that I get a cosine, and if I integrate
that I get back to the sine. They told me those things too.

[Eeyore]

Ian Iveson wrote:

the double integrator was removed from the feedback loop, so
that the accelerometer output was fed directly into the diff-amp.
Justification for this comes from the fact that , and from fourier
theory, which states that any given wave form can be reproduced as the
sum of a series of sine waves.

There is nothing in common in any way at all betwen a Fourier series and
integration (as in chalk and cheese).

The above idea is therefore based on a wholly flawed idea.

Graham

[Ian Iveson]

No I didn't write that. Arny posted it.

You know nothing. Shoo.


"Eeyore" wrote in message
...


Ian Iveson wrote:

the double integrator was removed from the feedback loop, so
that the accelerometer output was fed directly into the diff-amp.
Justification for this comes from the fact that , and from fourier
theory, which states that any given wave form can be reproduced as
the
sum of a series of sine waves.

There is nothing in common in any way at all betwen a Fourier series
and
integration (as in chalk and cheese).

The above idea is therefore based on a wholly flawed idea.

Graham