George (on Topic Only)

Actually making a basket is harder than you might think. DD went through at least 3 baskets (after tooling them up) making their "Beast" 18" woofer. The hard part is making something that won't crack under the stress that these new woofers are capable of.

 

Everything's tough if you don't have the right tools... We use Rhino for modeling, and once the basic "look" is finished, import into Wildfire (FEA package) for refinement/tweaking to ensure strength. Haven't had a failure of mechanical parts yet!

Dan Wiggins
Adire Audio

 

any input on this?!

eric

 

<---In envy. We don't have cool tools like that.

 

how linear would spandex be at excursion? :blink:

eric

 

any input on this?!

eric [/b][/quote]
Eric,

You're correct; of course, you effectively have a lot more motor force, too (4 motors with a "X" BL versus a single motor with "X" BL). So all things are no longer equal...

Dan Wiggins
Adire Audio

 

that actually occurred to me...4X Bl, 4x Mms, and a bunch of other stuff...

what this makes me wonder about now, is what do you have to give up to make the 15 have the same output in the same box as the 4 10's...are we talking 4X Bl (of the original driver), 4x Mms, and what else? the reason i ask, is because the 4 10's have 9 dB more output iirc as the single 15 due to powerhandling and by doubling Sd.

i really am trying to understand this

eric

 

How could a single 15" be as loud as four 10" subs?
It really boils primarily down to displacement, at a very high level...

But obviously that's discounting the type of power that's required to reach that excursion.

And it also is discounting the enclosure that's used to reach that excursion level, which has not only it's impact on how loud the actual output is (not to mention, what frequencies it's loudest at).

And the enclosure impacts how much power is required to reach a given excursion level.

Makes it difficult to say.
You can't really even say "given the same type of enclosure", because in the scope of two different drivers... what does that really mean?

 

sorry, this is taking us back a few pages... BUT we were talking about the Dumax & Klippel tests used to measure speakers,
WELL, here's a pic of a dumax test taking place. Thanks go out to Member02 for showing me the pic(and he's hosting it)

http://www.shafferflooring.com/dumaxbig.jpg

EDIT: well actually the test isn't taking place... but they are preparing for a test.

 

Wow..... I only read the first page and was duly impressed. I havent been on here for awhile, but if there is good discussion like this, I may stick around!!!


I havent really read the entire thread, but for my addition, I would like to discuss the paper Adire has published on their site about "woofer speed":
http://www.adireaudio.com/tech_papers/woofer_speed.htm

Hopefully dan will respond.


The paper attempts to remove cone mass as a factor to "woofer speed" and thus transient response, and instead point the blame on inductance. I say the paper is wrong on te cone mass, but I agree with the inductance theory. Adire makes an assumption that my high school math teacher would have smacked me across the head for.

In the paper, they say that BL*i = m*a and this is correct. However they make an assumption that you can drop both BL and m because they are constants..... but that is a BAD assumption. They are only constants on the same exact driver. Mass and BL vary from driver to driver.... so in fact:
(BL/m)*i=a

Lets take an example of two drivers...... Both have the same BL(say 20 newtons/ampere), but one has half the massas the other (say .2 kg vs .1 kg)

Now, we put the same current to both drivers... say 10 amps
Driver 1: a = 20*(20/.2) = 200 m/s^2

Driver 2: a = 20*(20/.1) = 400 m/s^2

The lighter cone driver accellerates TWICE as fast.


I was just wondering what Dans take on this is.....



Keep up the good thread guys.

 

Well it isn't acceleration that is the issue, but the change in acceleration. Heck, acceleration can be different for the same driver by changing frequency. The reason why acceleration is slower in one instance within your equations is because output is lower in one of those drivers. In other words the cone isn't moving as far, so acceleration isn't quite as high.

But what effects the CHANGE in acceleration? That would be current changes. What prevents current changes? That would be inductance.

 

hmmm... ok then CHANGE in accelleration. That would be da/dt or the differential of the above equation
so differentiating (BL/m)*i = a gives:
(BL/m)di/dt=da/dt
or the change in accelleration with respect to time is equal to the change in current with respect to time times a scalar (BL/m). Again, raise M change the the rate of change of accelleration with respect to time. Different way to look at it with the exact same results. Both BL and mass (as well as inductance) dictate how fast a driver can respond to a change in current.

 

Very good argument.

remember that it is the change in acceleration within the same driver. Once you build the driver, mass stays the same. So therefore mass isn't going to effect the change in acceleration since it is a constant. Current isn't constant, so it is the item that effects transient response.


I think I will go do some more studying on this. My above statement could be wrong in that mass may effect it slightly, but not as much as inductance. I will check on this.

 

Hey, just wanted to point you to a link that I have from a trip I took to DLC labs(where the DUMAX machine is), I've got pictures of one of my prototypes being DUMAXed, and we also DUMAXed a Koda 6.5" little guy:
http://www.betteraudio.com/geolemon/DLC
And there's a few other pics that are fun to look at. B)

I also had the honor of meeting Dr. Klippel at CES this year (he gave me a personal tour of the Klippel distortion Analyzer):
http://www.teamcaf.org/geolemon/CES2004/WKlippel.JPG
That's it, off to the left of the picture... (and me to the right :blush: ) but it's really more of a pic of the rig that can be used with the Klippel Distortion Analyzer... it's a modular system, that starts out remarkably reasonably priced ($8K-$12K), and simply requires a card that plugs into a computer, and some software. That pic shows the rig that holds the driver and a measurement laser, for an add-on component ($2K) to the Klippel system.

Hope that helps! B)

 

Definitely do stick around!

We've been striving to keep this place a more open, more diverse, yet more mature (less flaming) forum.

And we've had remarkable growth lately... this forum has decent traffic now, which is nice. Not CAF speed, but you might be surprised. B)

Many of us around here know you anyway, so it's all good.

 

And to bring this thread back on topic (something that we also like to do around here )...

Very good points, but only if your calculations are correct, obviously.
What formula are you using?
I question simply because you are referencing this formula immediately preceeding:
(BL/m)*i=a
...but that's not what you appear to be calculating.
Or possibly you simply used 20a rather than 10a in your calculation, you meant to use a BL of 20 and a current of 10a?

Calculation works out correct either way, just a little confusing to look at and extract from.

Now, I believe the point of the paper, and the statements saying that you can drop BL and mass were in the scope of a single woofer, a comparison of say, a given woofer, and then that exact same woofer with a heavier cone.

To use an Adire-specific example... the standard Brahma, and then the one that they provided to Scottie Johnson with a beefier cone, but was identical otherwise.

 

Wrong. BL and mass scale the amplitude of the change, but the change still begins at the time dictated by the inductance.

Inductance affects how fast I can change. BL and mass scale how much acceleration you get for a current. But they do not limit what currents you can put into the driver. Do the derivative, and what changes with respect to time? Current and acceleration. Mass and BL are purely scalars, and as such cannot be a time-variant quantity (they don't change with time). Thus they cannot affect transient response.

As far as your example, we see that we need less current for a given acceleration. But that is NOT transient response! Transient response is the rate of change of acceleration - the Jerk force (measured in units of Stapps). In your example, you need less current to get a given acceleration, but what limits the rate of change of the acceleration? Well, what limits the rate of change of current? Hint: it's not mass or BL...

If anything, that old "acceleration factor" misnomer of BL/m should be called an efficiency factor, because that's what it really is (BL and mass being the only physical parameters that set efficiency). Mass and BL are strictly an efficiency thing - inductance is a transient thing.

Dan Wiggins
Adire Audio

 

hi dan!

Yea, on another forum I was discussing the same topic and found out where my mathematical probleem was.

I was thinking of inductance as being constant, related by :
L = r^2 X N^2 / ( 9r + 10len )

where:
L = inductance in uH
r = coil radius in inches
N = number of turns
len = length of the coil in inches

but once I saw that it varied with excursion, it all made sense.... It still seems counter intuitive, and Im still in a search to disprove you guys

 

And hence the interesting thread on CAF you started...

Discussing the differences of opinion between Dr. Klippel and David Clark, regarding whether inductance should be included as a factor in determining Xmax.

I like the concept, but I think I like it best outside the scope of Xmax...
I like it under some new set of threshholds, like "Xsq" or something...

I think of Xmax as something that is more or less factors that relate to the speaker's ability to simply reach a certain excursion level linearly...
And it makes sense in that spec that it's the shortest of the two... Xmag or Xsus.

I'd love to see a spec that was separate, SQ influencing factors relative to excursion! That would be very nice.
Similar to how Xmax is derived from Xsus and Xmag...
"Xsq" (or whatever it would be) could be derived from some threshholds that are known more to affect SQ...
BL and even Kms could obviously be factors here...
But also inductance threshholds...
Cone/dustcap breakup threshholds...
Wind noise from the basket or pole vent threshholds..
etc.

I think it makes sense as a separate spec, personally (and I'd love to see it!)

Interesting to hear of their differences of opinion though!

 

I thought they had one called "THD". Of course it really isn't included in speaker specs.

But I like your idea Geo. Maybe have included on these testing reports (Kippel, DUMAX, Adire's) a breakup of even and odd order THD including average over the intended bandwidth and graphs included as well. Also maybe have those graphs take into account how audible those distortion levels are to a human (isn't that called the munsoon effect or something?). I also like the idea of having a breakup of each item you mention above in how much even and odd order distortion they add.