Hello.

I have some code that I wrote several years ago that lets me use a sound
card to accurately measure freq. vs. voltage. When normalized with a
resister, it can give me freq. vs. impedance for a speaker or a portion of a
crossover; anything reactive. I can also measure phase angle, independent of
voltage. So with this tool, I can measure many of the relevant TS parameters
and using the delta mass technique I can measure Vas.

My questions is this: Since the concept of mass doesn't involve any quality
or losses, like a box does, isn't it reasonable that, if done right, delta
mass would yield a much more accurate measure of Vas than the sealed box
method? And further, wouldn't using the sealed box measuring technique
backwards yield a better understanding of how big the box really is? In
other words, I can simply calculate H x D x W, but that might not be the
effective volume of the box. The shift in the resonant freq. of the woofer
in the box is perhaps more relevant.

My next question is this: Lets say the box measured above is measure as a
sealed box but intended to be ported. Now if I have the ability to measure
freq. vs. Z, can I use this to measure a change in the response that would
let me zero in on the exact tuning freq. for the box (port length)?

I have two completely different speaker design projects going at the same
time! I just got a delivery of 810 drivers today! I will be putting together
a website about both projects soon and I will post it here.

http://www.akrobiz.com/speakers/

Thanks!

James. )

James Lehman wrote:
My questions is this: Since the concept of mass doesn't involve any quality
or losses, like a box does, isn't it reasonable that, if done right, delta
mass would yield a much more accurate measure of Vas than the sealed box
method?

Generally, it is EASIER to get more accurate results using
the delta mass method. But the method described by Small does
account for simply absorbitive losses since it takes into
account both the new resonant frequency and the new Q at
resonance.

The problem comes in that the problems associated with the
sealed box method are not simply loss (i.e., energy dissipative)
issues. If they were, they could be accounted for. A leak, for
example, could also have a small reactive (acoustic inertance,
specifically). THat can complicate things in ways that are
difficult to account for.

However, all that being said, over a span of several thousand
friversm a fair number of which obtained the T/S parameters by
both methods, the results were within what I would expect
to be reasonable experimental tolerances.

The conclusion I would draw from the data I have before me is
that if appropriate care is taken in conducting the tests and
measuring the data, we don't see discrepancies that would
suggest that the sealed box method has systematic errors.

And further, wouldn't using the sealed box measuring technique
backwards yield a better understanding of how big the box really is? In
other words, I can simply calculate H x D x W, but that might not be the
effective volume of the box. The shift in the resonant freq. of the woofer
in the box is perhaps more relevant.

No, because it does not account for the fact that the effective
mass load on the driver COULD change.

My next question is this: Lets say the box measured above is
measure as a sealed box but intended to be ported. Now if I
have the ability to measure freq. vs. Z,

You'd have to to measure the T/S parameters anyway.

can I use this to
measure a change in the response that would let me zero in
on the exact tuning freq. for the box (port length)?

The "exact" tuning of a reflex not only involve port length,
but it involves enclosure volume as well.

But, that being said, the answer is most definitely yes. It
is certainly quite valid to use the impedance curve to verify
proper alignment of a system, whether it's a sealed or vented
or passive radiator system. It's a method I use all the time:
simulation is used to generate a proedicted impedance curve,
how close does the final result match the design target? The
impedance curve not only tells you how close you got, it can
tell you why you missed.

I have two completely different speaker design projects
going at the same time! I just got a delivery of 810 drivers
today! I will be putting together a website about both
projects soon and I will post it here.

Bah! There have been times when I had to verify that many
drivers a day, for a couple of weeks at a time. Great fun
it wasn't.

And further, wouldn't using the sealed box measuring technique
backwards yield a better understanding of how big the box really is? In
other words, I can simply calculate H x D x W, but that might not be the
effective volume of the box. The shift in the resonant freq. of the
woofer
in the box is perhaps more relevant.

No, because it does not account for the fact that the effective
mass load on the driver COULD change.

What? How?



My next question is this: Lets say the box measured above is
measure as a sealed box but intended to be ported. Now if I
have the ability to measure freq. vs. Z,

You'd have to to measure the T/S parameters anyway.


Of course I would! I'm not looking for any shortcuts here. I'm looking for
MORE things to measure.

The "exact" tuning of a reflex not only involve port length,
but it involves enclosure volume as well.

I understand the box and port together are a Helmholtz resonator; completely
independent of the woofer mounted in the box. But the woofer has to match
both the volume of the box and the tuning freq.



But, that being said, the answer is most definitely yes. It
is certainly quite valid to use the impedance curve to verify
proper alignment of a system, whether it's a sealed or vented
or passive radiator system. It's a method I use all the time:
simulation is used to generate a proedicted impedance curve,
how close does the final result match the design target? The
impedance curve not only tells you how close you got, it can
tell you why you missed.

How? What do you look for?



I have two completely different speaker design projects
going at the same time! I just got a delivery of 810 drivers
today! I will be putting together a website about both
projects soon and I will post it here.

Bah! There have been times when I had to verify that many
drivers a day, for a couple of weeks at a time. Great fun
it wasn't.


Well... I'm not working for someone else. Maybe that's why I love what I do!

James. )

James Lehman wrote:
And further, wouldn't using the sealed box measuring technique
backwards yield a better understanding of how big the box really is? In
other words, I can simply calculate H x D x W, but that might not be the
effective volume of the box. The shift in the resonant freq. of the
woofer
in the box is perhaps more relevant.

No, because it does not account for the fact that the effective
mass load on the driver COULD change.

What? How?

The total effective mass on a driver includes not only
the mechanical mass of the voice coil/diaphragm, etc., but
also the mass-reactive portion of the radiation impedance,
which is dependent upon how it is loaded (baffle size, for
one). The incremental amount of mass is not huge, on the
order of a few percent, but it can be significant. It is why,
for example, all of my driver measurements, even mass-loaded
T/S measurements, are done in standard sized-baffles.

Take a look at the radiation impedance in the piston band.
It rises as a function of frequency and is both reactive
and resistive (both are increasing as a function of frequency,
and noth dependent on the baffling).

This is why, often, you will see, especially in some of Small's
articles on vented system analysis, he uses the parameter Fsb
for driver fundamental resonance, not Fs. The difference is that
Fsb represented the baffle/enclosure loaded resonant frequency,
While Fs is the "free air" resonance, i.e., no baffling of any
kind.

My next question is this: Lets say the box measured above is
measure as a sealed box but intended to be ported. Now if I
have the ability to measure freq. vs. Z,

You'd have to to measure the T/S parameters anyway.

Of course I would! I'm not looking for any shortcuts here. I'm looking for
MORE things to measure.

The "exact" tuning of a reflex not only involve port length,
but it involves enclosure volume as well.

I understand the box and port together are a Helmholtz resonator;
completely independent of the woofer mounted in the box.

Well, no, not "completely independent." They are coupled forced
harmonic oscillators, after all.

But the woofer has to match
both the volume of the box and the tuning freq.

Well, not exactly. If what you mean (your statement parses
ambiguously) is that the woofer Vas needs to match the enclosure
Vb, and the woofer Fs has to match the box Fb, then that is true
for only one VERY specific alignment: the so-called lossless
B4 maximally-flat system transfer function, and that requires
a VERY specific set of T/S parameters.

If, instead, you mean that for a target system transfer function
based on the driver's T/S parameters, you need a specific box
volume and tuning frequency, then yes.

But, that being said, the answer is most definitely yes. It
is certainly quite valid to use the impedance curve to verify
proper alignment of a system, whether it's a sealed or vented
or passive radiator system. It's a method I use all the time:
simulation is used to generate a proedicted impedance curve,
how close does the final result match the design target? The
impedance curve not only tells you how close you got, it can
tell you why you missed.

How? What do you look for?

Well, without getting into the very nitty gritty details,
the impedance curve is merely one form of the system transfer
function. As such, if you can derive the impedance curve from
the driver and enclosure parameters (you can, indeed), and since
the electrical impedance function is a minimum-phase function,
then if you have the impedance curve, you can derive the driver
and enclosure parameters. Basically, the driver/enclosure
parameter to electrical impedance is symmetrically reversable.
I think it was Leach who described this process in detail about
25 years ago in JAES.

I have two completely different speaker design projects
going at the same time! I just got a delivery of 810 drivers
today! I will be putting together a website about both
projects soon and I will post it here.

Bah! There have been times when I had to verify that many
drivers a day, for a couple of weeks at a time. Great fun
it wasn't.
Well... I'm not working for someone else. Maybe that's why I
love what I do!

Neither was I, at least in the employee sense of the term. One
of my clients at the time had taken delivery of several thousand
custom drivers manufactured by an OEM company that honestly
believed that if a surround LOOKED the same, it WAS the same.
About 40% of the drivers were seriously out of spec, resulting
in an Vas that was about 30% too low and an Fs that was thus
about 15% too high. Since the client was essentially an
"integrator" (they were assembling the drivers and crossovers
into cabinets they were building for a third party), they had
relatively little expertise in measuring and qualifying the
drivers, so it was left to me, though atv about $75/hour at
the time (which was eventually coughed up by the driver manu-
facturer, thoroughly erasing ANY profit they hoped to make
by substituing surrounds instead of buying the right ones), I
suppose I shouldn;t complain. But I do anyway.