Cross over question

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

Dumb question here . . . The cross over design I have says in a note on
the tweeter circuit that the value of one of the resisters can be
changed to adjust the tweeter loudness. If I increase the resister
value does that make the tweeter louder?
Thanks, Jim
--
When emailing to me change pactell.net to pacbell.net. Thanks

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

"Jim H" <jb@pactell.net> wrote in message
news:nGGye.796$6%2.355@newssvr21.news.prodigy.com...
> Dumb question here . . . The cross over design I have says in a note on
> the tweeter circuit that the value of one of the resisters can be
> changed to adjust the tweeter loudness. If I increase the resister
> value does that make the tweeter louder?

You need to provide the circuit. Sometimes, the circuit will have a
resistor in series with the tweeter, and another resistor in parallel to
that resistor/tweeter combination. Increasing one of the resistor values
will make the tweeter louder, increasing the other will make it quieter.

Tim

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

James Lehman wrote:
> What does that mean? Any dynamic magnet / motor driven speaker has a
> suspension and the whole moving system has at least one major resonance.

Actually this is true of every loudspeaker, at least every
one that has mass and some compliance (yeah, that'd be every one).
Electrostatics have resonances, piezo drivers have resonances,
and all exhibit impedance peaks that hese resonances

> All
> sorts of weird things happen when a signal within this region is fed into
> the speaker. The peak impedance of the driver is WAY higher than nominal

No, the impedance MAY be way higher than nromal. It is not a given
that it MUST be so.

> and
> the phase of the resulting output of the driver, compared to the signal
> being fed into the driver, will shift within this region according to the
> order of the crossover network. The lower side of the resonance will
> actually be time shifted from the upper side.

Trying to seperate the signal separately into its phase and amplitude,
as your doing here, is not useful. Consider the signal as a complex
signal, and the analysis becomes more complete and way easier.

> Most people don't think about the box as being part of the crossover, but it
> is! The box is the high-pass filter for the woofer. A sealed box is much
> like a capacitor in series with the woofer. It imposes a -6dB slope from Fb
> (resulting resonance of the woofer in the box) down.

False, absolutely false.

ALL mechanically resonant systems will exhibit an asymtotic
slope of 12 dB/octave below resonance. The addition of a sealed
box to a driver will NOT "impose a -6dB slope." The driver itself
already has the 12 dB slope inherent because it is a 2nd order
mechnaically resonant system. The addition of the box simply
changes the totoal system compliance, raising the frequency of
resonance. It will still, afterwards, have a 12 dB/octave slope.

>A Ported box is sort of
> like a capacitor and a coil. It has a -12dB slope.

False again.

The driver, as in the case of either performance in free air
or in a sealed box, still has a 12 dB/octave slope. Placing
the driver in a vented box adds two more orders to the system
response.

> One thing you NEVER want is resistance in series with a woofer.

False again.

The total series resistance is what determines the major portion
of the driver damping at resonance. Having no series resistance
means the driver has infinite damping. Having too small a series
resistance means the driver is overdamped. The right amount of
resistance is the right amount of resistance, and that's never 0.

> Unfortunately, it is unavoidable. There is plenty of it in the speaker wire
> and coil that might be in series with your woofer.

Unformtunately, you have ignored the most significant source of
series resistance, and that's simply the resistance of the voice
coil.

> You just have to deal with that.

And a good designer WILL deal with it, knowing what the target
response of the system needs to be, and adjusting all resistances
accordingly. It is not, for example, at all unreasonable or incorrect
to take a woofer that is electrically overdamped and wind a series
crossover inductor with sufficiently small gauge wire to add
enough DC resistance to achieve the target system damping. It is
a prefetcly valid tehcnicla approach when you are stuck with a driver
whose total damping is too high.

> Drivers to be used above the woofer need to be at least as or more efficient
> than the woofer. If a tweeter is more efficient than a woofer, it may be
> attenuated with a resister in series with it. But putting a resister in
> series with the tweeter will create a device that has the sum of the
> resistance and the impedance of the tweeter. So another resister needs to be
> put across this device in parallel to bring down the total impedance of the
> device to match what the tweeter alone would have been. It is possible to
> design a crosser that doesn't need this extra resister, simply by
> calculating the parts formulae using the sum of the series resister and the
> impedance of the tweeter, but this higher impedance in the tweeter range of
> the total system will show all the way back to the amplifier.

So? The amplifiers we use today are essentially voltage sources,
so it's not really much of an issue, is it?

> It is usually
> best to try to maintain as flat of an impedance curve to the amp as
> possible.

Then why not correct for the WAY higher impedance peak at resonance?

> Take a look at using an L-Pad.

An L-pad will only work proeprly IFF:

1. The load impedance matches the requirements of the L-pad,
2. The load impedance is resistive.

Few tweeters meet these requirements.

> Another aspect of crossover design that is often missed is compensation for
> the rise in impedance cause by inductive reactance of the voice coil. This
> has most effect in the midrange to tweeter crossover region. A Zobel network
> needs to be placed in parallel with the midrange to flatten out the
> resulting impedance. Otherwise, the crossover will not produce the nice
> symmetrical roll-offs you want.

Now, it has little to do with "nice symmetrical rolloffs." It
has everything to do with the fact that it is impossible to
achieve controlled response of any sort using passive ladder network
filters into non-resistive loads.

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

James Lehman wrote:
> I hadn't forgotten that this is your Internet, dpierce@cartchunk.org! And
> once again you make all sorts of arguments about what I didn't write; not
> what I actually did write. Some things just never change.

Let's in fact see what you DID write.
> > > Most people don't think about the box as being part of the crossover,
> but it
> > > is! The box is the high-pass filter for the woofer. A sealed box is much
> > > like a capacitor in series with the woofer. It imposes a -6dB slope from
> Fb
> > > (resulting resonance of the woofer in the box) down.
> >
> > False, absolutely false.
>
> TRUE! A sealed cabinet does exhibit the properties of a non-polarized
> capacitor. Any displacement of the cone stores energy in the pressure
> difference between the air inside and outside the box. In a perfectly sealed
> box with no air flow through the dust cap of the woofer, there is no other
> way to equalize the pressure, except to move the cone back to the center of
> its throw.

You stated, and I quote you exactly:

"It imposes a -6dB slope from Fb (resulting resonance of the
woofer in the box) down."

Do you deny saying that? If not, then if you are capable of reading,
it is precisely THIS point I was addressing. And that point is
wrong.

The imposition of additional system stiffness DOES NOT IMPOSE A
-6dB SLOPE FROM Fb. That is the point which you got wrong. The
natural rolloff of the woofer is already 12 dB per octave, the
addition system stuffness does not, as you clearely state, impose
a 6 dB roloff, it simply changes the resonant frequency of the
system. The slope is STILL asymtotic to 12 dB per octave

Now, the statement that YOU made was:

"It imposes a -6dB slope from Fb (resulting resonance of the
woofer in the box) down."

And it is that statement, which is your statement, that I challenged.

> > ALL mechanically resonant systems will exhibit an asymtotic
> > slope of 12 dB/octave below resonance. The addition of a sealed
> > box to a driver will NOT "impose a -6dB slope." The driver itself
> > already has the 12 dB slope inherent because it is a 2nd order
> > mechnaically resonant system. The addition of the box simply
> > changes the totoal system compliance, raising the frequency of
> > resonance. It will still, afterwards, have a 12 dB/octave slope.
>
> That's why I used the phrase "It imposes a -6dB slope". This is in addition
> to the dynamic characteristic of the naked woofer. Again, what I didn't
> say...

Let's try this once again: IT DOES NOT IMPOSE A -6DB SLOPE OR
ANY OTHER KIND OF SLOPE on the system. A driver with no enclosure
is a 2nd order mechanical system with a rolloff below resonance
which is asymtotice to 12 dB per octave. The very same driver in
a selaed box IS STILL A 2nd OREDER mechanical system with a rolloff
asymtotic to 12 dB per octave. The presence of the additional system
stiffness DOES NOT, as you claim, impose a -6dB slope.

> > >A Ported box is sort of
> > > like a capacitor and a coil. It has a -12dB slope.
> >
> > False again.
> TRUE! Once again the air inside the box can and does temporarily store
> energy as a pressure difference, much like a capacitor. But this time there
> is also a reactive circuit to ground; a hole in the box; much like a coil,
> allowing the pressure difference to dissipate. So displacing a woofer cone
> in a ported box does not cause a lasting pressure difference in the box.

Because your simplistic model of "displacing the woofer cone
in a ported box [that] does not cause a lasting pressure difference"
treats the port not as an acoustic inertance, which is is, but as
a dissipative element, WHICH IT IS NOT.

> > The driver, as in the case of either performance in free air
> > or in a sealed box, still has a 12 dB/octave slope. Placing
> > the driver in a vented box adds two more orders to the system
> > response.
> Two more orders... As in a cap and a coil... So you agree!

No, I do not.

> > > One thing you NEVER want is resistance in series with a woofer.
> >
> > False again.
> >
> > The total series resistance is what determines the major portion
> > of the driver damping at resonance. Having no series resistance
> > means the driver has infinite damping. Having too small a series
> > resistance means the driver is overdamped. The right amount of
> > resistance is the right amount of resistance, and that's never 0.
> >
> > > Unfortunately, it is unavoidable. There is plenty of it in the speaker
> wire
> > > and coil that might be in series with your woofer.
> >
> > Unformtunately, you have ignored the most significant source of
> > series resistance, and that's simply the resistance of the voice
> > coil.
> A woofer is a complex dynamic system. DC resistance in the coil is an
> absolutely necessary and controlled thing that has everything to do with the
> power of the piston motor, length and diameter of the wire in the coil, mass
> of the coil, etc, etc... "A woofer" et al, is a resister, among other
> things, and there is NO good reason to put another plain resister in series
> with it ever.

I would suggest that you might ewant to broaden your vision
from your clearly very narrow perspective. Having, in fact,
worked with many manufacturers of both drivers and systems, and
been confronted, at every case, of having to develop a system
based on non-optimal components. it is OFTEN the case where
a systems manufacturer does not have available the EXACT driver
needed: the design budget simply does not allow for customization.
Further consider the fact that for such manufacturers, cabinet volume
is predetermined and thus is not design variable. In such cases, it
is not uncommon to tune available drivers to suit the enclosure, and
not the other way around.

TO claim that such a technique is invalid, as you essentially do,
is to exhibit the most narrow-minded of viewpoints towards system
design and optimization.

Curiously, this exact technique is recommend by Small, Thiele, Benson
and others in the field. Perhaps you might share with us the wisdom
that you have that trumps theirs.

> > > You just have to deal with that.
> >
> > And a good designer WILL deal with it, knowing what the target
> > response of the system needs to be, and adjusting all resistances
> > accordingly. It is not, for example, at all unreasonable or incorrect
> > to take a woofer that is electrically overdamped and wind a series
> > crossover inductor with sufficiently small gauge wire to add
> > enough DC resistance to achieve the target system damping. It is
> > a prefetcly valid tehcnicla approach when you are stuck with a driver
> > whose total damping is too high.
>
> Either fix the box or get a better woofer.

You clearly do not understand system optimization and engineering.
The box is often dictated by marketing consideration or other
engineering constraints. Woofers parameters are NOT available in
any arbitrayr combinations. One may be able to get close, but if
the box size is constrained and the woofer is electrically overdamped,
underdamping it by adding resistance is a perfectly valid way of
getting to the end point.

And let's just say we COULD get a "better woofer", one with lower
electrical damping. Let's say we could magically reduce the Bl
product by the right amount (a tough job, since magnets and magnet
circuit components are not available in arbitrary sizes). The lowered
Bl product needed to reduce the electrical damping and raise the
electrical Q qill result in, for example, a driver efficiency that
WILL BE EXACTLY THE SAME AS TAKING THE ORIGINAL DRIVER AND ADDING
ENOUGH RESISTANCE TO RAISE THE Q.

The difference being that a $0.50 resistor for 5,000 woofers on
a production run is often cheaper, especially time-wise, than
paying for custom tooling of a driver whose parameter are exact,
on the possibly mistaken assumption it's even possible. (Because
magnets are only available in discrete sizes, because front plates
are only available in 2 or 3 standard thicknesses, because wire
diameters are only available in discrete sizes, and because voice
coils can only be wound with integral numbers of layers, parameters
are NOT available over a continuum: they are, in fact, fairly
pretty quantized.)

Now, that's the REALITY of designing loudspeakers.

> > So? The amplifiers we use today are essentially voltage sources,
> > so it's not really much of an issue, is it?
>
> "The amplifiers we use today"... What are those? Who's we? I'd like to think
> that there may be any sort of amplifier out there in use. Every amplifier
> ever made for general compatibility with the rest of the audio market was
> designed to be stable while operating into some sort of nominal load; most
> likely a big fat resister. What about tube amps with impedance matching
> output transformers?

With the exception of pathologically designed single-ended tube
amplifiers, the vast majority of amplifiers in use for audio purposes
act essentially as voltage sources when loaded with the vast majority
of available loudspeakers. And that includes thge vast majority of
tube amplifiers with output transformers. You may know of some
exceptions. Fine, but they are rare.

> > > It is usually
> > > best to try to maintain as flat of an impedance curve to the amp as
> > > possible.
> >
> > Then why not correct for the WAY higher impedance peak at resonance?
>
> Why not! It's a very easy thing to do. It is very similar to a Zobel
> network. Can and has been done. Plus, getting back to "the box is a part of
> the crossover" statement, The box does a lot to alter the impedance curve of
> the woofer.

Not once you get out of the realm where the motional impedance
dominates, it doesn't. An 8" driver in a reflex enclosure tuned
to 40 Hz has no significant effect on the impedance of the driver
above a couple of hundred Hz.

> The only way to really know the whole story is to measure the
> total system impedance after it is fully assembled... and even
> sitting in its eventual listening position in the room.

Oh, please. Unless you can actually SHOW us specific repeatable
examples of how the listening position affects the impedance
sufficiently to change the crossover response, and I have measured
several thousand speakers in situ, we can chalk this claim up to
uininformed specualtion.

> > > Take a look at using an L-Pad.
> >
> > An L-pad will only work proeprly IFF:
> >
> > 1. The load impedance matches the requirements of the L-pad,
> > 2. The load impedance is resistive.
> >
> > Few tweeters meet these requirements.
>
> The schematic diagram of an L-Pad will show you the series and the parallel
> resisters and the way they relate to the driver. Any driver that is rated as
> an 8 ohm device will most likely be very close to 8 ohms within its nominal
> operating range.

Fine, hook a 4-ohm L-pad to a Vifa XT25TG30, which is rated at
a nominal 4 ohms, and try to explain awya the gross peak that
occurs at 750 Hz as a result.

> The DC resistance of the coil is most of the value; usually
> more than 6 ohms of it.

Only if that level of attenuation is required. For lower attenuation,
the series resistance will be lower.

Now, once again, all the textbooks on L- or T-pad attenuators
state quite clearly that the efficacy is highly dependent on
two factors:

1. The load impedance IS the design impedance,

2. The load impedance is resistive.

Again, if you have wisdom that trumps many decades of engineering
expertise in the field, please do share it with us.

> There's only about a bazillion tweeters out there and at least
> six or eight of them are half way decent.

And what if the one you need requires sufficient attenuation and
does not provide a resistive load?

> > Now, it has little to do with "nice symmetrical rolloffs." It
> > has everything to do with the fact that it is impossible to
> > achieve controlled response of any sort using passive ladder network
> > filters into non-resistive loads.
>
> Impossible! Oh my god! Then why do we even bother?

Read what was written yourself. It is the entire point behind complex
conjugate compensation.

> I guess its about doing
> the best we can with what we really have to work with.

And I would humbly suggest that your comments suggest you don't
have a lot to work with.

> Spell checkers are quite useful things too.

And irrelevant when you insist on holding naieve, untenable technical
viewpoints.

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

James Lehman wrote:
> I think it's really something that anything that I wrote would
> cause you to insult me.

If you maen that testing your technical assertions is "insulting,"
then I might suggest that you not make such assertions. You need
to learn the difference between challenging technical assertions
and your personal space. They are not the same at all.

> I'm an artist and I try to use as much engineering and design as I can to
> make my art better. It is a constructive effort that brings good things into
> existence. I learn from it and I like to share my experience.

A congratualte you on your efforts.

However, what relevance does that have to the fact that you made
a number of technical assertuions that were, in fact, provably
wrong. For example, your assertion that a sealed box imposes
a -6dB slope on a woofer is an assertion which is not made by
ANYONE else in the field. Not Thiele, not Small, not Benson,
not Beranek or Kinsler or Frey or Coppens or anyone else makes
such an assertion. Nor does there exist any body of technical
work in the field that even hints at such.

Yet, you consider any correction to your incorrect technical
assertion as insults? You need to seriously recalibrate tourself,
my friend, you;re way out of line.

If you make incorrect assertion of this nature, do you think they
should stand unchallenged? What is it about your assertions then
that makes them unchallengable?

Kind sir, you made several technical assertions that are testable
and, when tested, proved to be incorrect. Deal with it. The rest of
us do by correcting what we think we know anmd not making that
particular set of mistake again. Any reaosn why you're exempt from
this?

Let's go back to just one of your statements: that a selaed box,
in your words, imposes a -6dB slope on the system. That statement,
all by itself is purely and simply wrong and illustrates that
wherever that notion came from, the source of that notion has a
fundamental misunderstanding of the basic principles behind loud-
speakers and mechanically resonant system.

Consider the following electrical analog if you will. You have a
2nd order high pass filter consisting of a series capacitor and
a shunt inductor going to a load. A SPICE model will unambiguously
illustrate the topology:

Vin 1 0 AC SIN 1.0 0.0
CSeries 1 2 280UF
LShunt 2 0 36MH
RLoad 2 0 8

This circuit has a cutoff of 50 Hz.

Now, put a second cap in parallel with the first. This is exactly
equivalent to the case of putting a driver in a sealed box, ie.
increasing the total system stifness. Then resulting filter looks
like:

Vin 1 0 AC SIN 1.0 0.0
CSeries 1 2 280UF
CSeries1 1 2 280UF
LShunt 2 0 36MH
RLoad 2 0 8

In this case, we assume that the "cabinet" has the same effective
acoustical compliance as the driver. The net resuklt is STILL a
second order high-pass, the slope is STILL 12dB per octave, and
the addition of the second capacitor has NOT "imposed a -6dB slope"
at a all.

Indeed, the result is EXACTLY equivalent to:

Vin 1 0 AC SIN 1.0 0.0
CSeries 1 2 140UF
LShunt 2 0 36MH
RLoad 2 0 8

Now, please show us, if you weill, how your claim that a sealed
box "imposes a -6dB slope" is technically valid and, if it is not,
why it does not need to be corrected.

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

"James Lehman" <james[remove]@akrobiz.com> wrote in message
news:C9CAe.17405$Tf5.16509@newsread1.mlpsca01.us.to.verio.net...
> That says more about you than it does about me!

What you incorrect posts and your childish attempts to criticise Dick, say
about YOU, is best left unsaid.

> I'm an artist and I try to use as much engineering and design as I can to
> make my art better. It is a constructive effort that brings good things
into
> existence. I learn from it and I like to share my experience.

Maybe you should confine sharing your experience to whatever you actually
have some clue about. Art maybe?
You should thank Dick profusely for taking his time to try and educate you.
Instead you abuse him.

Fortunately for the people who may otherwise be mislead, Dick continues to
offer his knowledge and experience to those who appreciate it.
(and even those like you who don't :-)

MrT.

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

"James Lehman" <james[remove]@akrobiz.com> wrote in message
news:C9CAe.17405$Tf5.16509@newsread1.mlpsca01.us.to.verio.net...

> I think it's really something that anything that I wrote would cause you
to
> insult me.
>
> That says more about you than it does about me!

He has some baffling ideas; he claimed you couldn't have two periods of
silence, one lasting one microsecond longer than the other.

Tim

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

Tim Martin wrote:
> <dpierce@cartchunk.org> wrote in message
> news:1121105349.164597.42070@o13g2000cwo.googlegroups.com...
>
> > ALL mechanically resonant systems will exhibit an asymtotic
> > slope of 12 dB/octave below resonance. The addition of a sealed
> > box to a driver will NOT "impose a -6dB slope." The driver itself
> > already has the 12 dB slope inherent because it is a 2nd order
> > mechnaically resonant system. The addition of the box simply
> > changes the totoal system compliance, raising the frequency of
> > resonance. It will still, afterwards, have a 12 dB/octave slope.
>
> This is wrong.

No, it's not. You completely failed to read and comprehend what
was written.

Indeed, ALL mechanically resonant systems exhibit asymtotic slope
of 12 dB per octave below resonance. The imposition of additional
resonant system, such as a reflex, will add more poles to the total
response, one pole pair per resonant system.

On the assumption we are discussing drivers operating in the piston
region, the requirements for flat response are maintained only for
systems whose volume velocity is independent of frequency, and that
condition holds for mechanical resonant system operating in the mass-
controlled region, which is above resonance. In that region,
acceleration is constant with frequency, velocity goes as the
reciprocal of frequency, and excursion goes as reciprocal of
the square of frequency. These conditions maintain the volume
velocity required for flat response.

Below resonance, the drive is operating in the stiffness controlled
region, where acceleration now goes as the square of frequency,
velocity foes directly as frequency, and excursion is constant.
The conditions required for flat response are no longer met in the
stiffness controlled region of operation, and the response MUST
as a consequence roll off at 12 dB octave.

These mechanical conditions are extant regardless of the baffling,
and impose the fundamental limits to how much power can be produced
by the system.

You really need to go staudy and understand mechanically resonant
systems before you make such assertions.

> Transmission line

False. This is myth unsupported by either theory or measurement.
Below line resonance, the response rolls off at a minimum of 12
dB per octave imposed by the driver mechanical properties as
described above and then approach 18 dB per octave due to the
zero in the transfer function imposed by simple cancellation.
There is no mechanicam by which one pole of the driver's 2-pole
response can be cancelled by a tyransmission line.

> and open baffle speakers have a -6dB/octave slope

False. This is myth unsupported by either theory or measurements.
Having a driver in an open baffle does not change the fundamentl
mecganical operation of the driver as described above. Cancelation
simply imposes a zero in the response at, essentially 0 frequency.
It may START to roll off at 6 dB per octave due to the cancellation
zero in the reposnse, IF the baffle size is sufficient;y small such
that the cancellation rolloff occurs above resonance, but the
rolloff rate will eventually be dominated by the mechanical rolloff
described above.

There is no mechanicam by which one pole of the driver's 2-pole
response can be cancelled by an open baffle.

> Infinite baffle have -12fB/octave slope

The baffle has NO rolloff rate. You are simply seeing the mechanical
rolloff of the driver with nothing else imposed on it.

> Bass reflex have -24dB/octave slope

The driver's fundamental rolloff rate itself is 12 dB/octave, still,
for the reasons described above.

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

"James Lehman" <james[remove]@akrobiz.com> wrote in message
news:Ix_De.20643$Tf5.13699@newsread1.mlpsca01.us.to.verio.net...
> I have a very "matter of fact" tone to a lot of what I say and write.

I think the gist of what you said is correct, it was just the details I
though needed correcting.

Take a concrete example ... suppose you're using the CDT Audio CDT-5 as a
mid-range unit. This has an fs of 80Hz, and a Vas of 4.5 litres. If we
install it in a 2-litre enclosure with no crossover, it will roll off at
12dB/octave from about 150Hz downwards.

So on the face of it, we can couple it with an LF driver with a 12dB/octave
150Hz low-pass filter and get a flattish overall response. If we want a
steeper filter, we can use a 24dB/octave 150Hz low pass filter coupled with
a 12dB/octave 150Hz high-pass filter for a flattish response.. And because
the electrical filters are different orders, they cause different delays,
resulting in a phase difference at the crossover frequency (for 12db/octave
coupled with 24 db/octave, the phase difference is 180 degrees, so we can
compensate by reversing the leads to one driver.).

At least, if I understood you correctly, this is what you were driving at.

Tim

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

James Lehman wrote:
> I was wrong about the idea that a sealed box imposes its own effect
> of a -6dB slope.

Yes, you were.

> But a sealed box DOES emulate a similar effect as putting a
> capacitor in series with the woofer or stiffening the suspension.

No, it does not. Putting a capacitor in series with a woofer
adds a third pole to the lower frequency response and thus
changes the asymtotic rolloff rate from 12 dB to 18 dB per
octave. Stiffening the suspension does NOT add a third pole
to the low frequency response. It simply raises the resonant
frequency and raises the Q at resonance. The system is STILL
2nd order and the asymtotic rolloff rate remains 12 dB per octave.

> Although, a large capacitor would prevent DC from ever
> getting to the woofer; one of the reasons why most power
> amps have output decoupling caps.

Most power amps DO NOT have output coupling capacitors. You would
do well not to assert your opinions as facts.

> Also, try putting a woofer with an already high Q in a tiny
> box and see what you get! A HUGE resonant peak way above the
> free air resonance of the woofer

If and only if the enclosure stiffness dominates. Take a high-Q
woofer with a high suspension stiffness and put it in the same
enclosure, and the system resonance is NOT "way above" free air
resonance.

> and a slope on the down side that crashes fast and hard.

Yet, if you examine the system, the asymtotic rolloff rate is
STILL 12 dB per octave.

> Pass band boxes are woofers with boxes behind and in front of the cone.
> That's a high pass and a low pass filter.

NO, let's try this once again. The high-pass filter is a direct
result of the fact that the driver is a 2nd order high pass
mechanically resonant system to begin with. The front side
acoustical cstiffness of the cabinet and the iniertance of the
port form a 2nd order acoustically resonant low-pass filter.
The result is a 4th order bandpass system.

And, should you taker the time and step back and analyze the
configuration, you will come to understand, one would hope,
that this is precisely what's happening in a reflex system as
well. The woofer is ALWAYS a second order high-pass. The enclosure
is a 2nd order low pass. If you look at the output of the
port in a reflex system, the result is a 4th order bandpass
function. Tuning a reflex system is essentially optimizing
the response of the direct output of the woofer, with its 2nd
order, preferably overdamped high pass response, and the
output of the port, a 4th order bandpass, such that the port
supplants the output of the driver at and near cutoff.

> There is no doubt that a ported box adds bass extension
> to the working system that exhibits a 180° phase shift.

Well, maybe in your view. But Messrs. Thiele, Small, Benson
and many others would cast great doubt o your assertion. As
would I.

Until you are prepared to examine the complex transfer function
of the woofer itself, including the acoustical load imposed by
the enclosure, and the complex transfer function of the port,
and be prepared to understand how the sum of the total volume
velocity of both contribute to the system output, you're going
to keep making the same incorrect assertions. That may be your
choice, but it is certainly yours to make.

A properly configured reflex system CANNOT "add bass extension."
The driver has to be able to make it down to the desired cutoff
frequency to begin with. The port's contribution to the volume
velocity does not "extend" the output of the woofer: it in fact
REPLACES the output of the woofer at and around the enclosure
resonant frequency.

The "bass extension" that many people assume is gained from a
reflex system is a result of a narrow and incomplete understanding
of the operations and requirements of such systems. A proper
understanding of the efficiency-bandwidth-volume relationship
is the only way to comprehend what is actually going on.

The advantage gained from using a reflex is NOT that it "adds
bass extension." It's that is capable of as much as a factor of
two increase in the efficiency-bandwidth-volume equation. Only
if you make two of those parameters the same, namely efficiency
and enclosure volume, does the bass extension hold, and only then
by a mere 1/3 octave. the relation being:

N0 = kn Vb F3^3

where kn is the efficiency constant, which depends upon system
order and alignment, and is about twice that for the best-case
reflex (B4 lossless alignment) as it is for the best-case sealed
box (C2 Qt=1.2 or so), Vb is the volume of the enclosure, F3
is the -3dB low end cutoff frequency, and n0 is the reference
efficiency.

How does this work?

It means, in this particular case, that for a given driver
efficiency and enclosure volume, that a reflex IS capable,
assuming the rest of the driver parameters are suitable,
of operating 1/3 octave lower than a sealed enclosure. But
the secret is NOT that the reflex somehow magically lowered
the bandwidth, it's because the driver that's suitable for
the reflex system MUST have a set of parameters that results
in its fundamental resonant frequency being lower. And that
means, almost always, for a given volume, having a similar
suspension stiffness, a higher cone mass and, to get that
efficiency back and thus lower the Qt raised by the higher
mass, a higher Bl product.

Taking a sealed cabinet and popping a port in it will not
magically lower the cutoff frequency, any more than it will
raise the efficiency of the system. This is because the
efficiency constant kn is dependent, as mentioned, on system
order an alignment. Take, for example, the optimal B4 reflex
alignment. Plug up the port, and you end up with an overdamped
sealed box system, but one with exactly the same efficiency.
Yes, the -3dB point is somewhat (but not very much) higher,
but the asymtotic rolloff rate, being now 12 dB per octave,
means a lower F10 in many cases, especially given the resulting
system is now overdamped. The approach of the response rolloff
to the asymtotic rate is delayed for some bit below cutoff.

Let's examine one last time the assertion you still insist
on making:

"a sealed box DOES emulate a similar effect as putting a
capacitor in series with the woofer or stiffening the
suspension."

In way, shape or form is this true. Changing the total system
stiffness, whether mechanically or acoustically, is not emulated,
not even approximately, by adding a capacitor in series with the
woofer. The former changes the resonant frequency and Q, but the
system ALWAYS remains a 2nd order high-pass. The latter transforms
the system from a 2nd order to a 3rd order high-pass.

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

James Lehman wrote:
> Exactly! I agree with you completely. I couldn't have said it any better
> myself. You are a genius. I am in such awe of your exquisite intelligence
> and infinite compassion. You are my hero! Golly, what a great guy! I think I
> will name my first child dpierce@cartchunk.org!

Another of the impassioned, germane, technically erudite,
brilliantly argued posts of the very kind we have come to
expect of you, James.