Electrical vs. Acoustic Slopes

[sdclc126]

Greetings All,

I've noticed in descriptions of several crossover designs, for example Zaph's BAMTM, that a 4th order acoustic slope was achieved with a 2nd order electrical construct.

My question is, is this a universal outcome, or is it dependent on a set of variables (and what are they)? And given this, what then would be the acoustic slope of a 4th order electrical construct, under the same circumstances? 6th order? 8th order? What would be the acoustic slope of the BAMTM crossover if it were actually constructed 4th order electrical? When does the acoustic slope actually equal the electrical, or does it ever?

Please help the crossover challenged!

[joe carrow]

A driver might have some rolloff without any crossover at all. The crossover adds to that. From what I've seen, the driver rolloff doesn't tend to be as much like a textbook straight line as an electrical filter. This means that adding electrical filters to the mechanical system will give you an approximate effect, and it will depend on the specifics to determine how it actually adds up.

If you want to get 4th order acoustic slopes from 4th order electrical filters, then you need a driver that's nice and linear (flat frequency response) for a good bandwidth above and below the crossover point- say 2 octaves on either side. Once the response is attenuated enough, it matters less if it starts to fall off faster. With 24 db/octave, one octave away from the crossover point the driver is down by 24 db, at two octaves it's 48 db- beyond that, who cares?

First order, on the other hand, is darn near impossible to do right. Do you see why?

[Lynn Olson]

Originally posted by sdclc126
Greetings All,

I've noticed in descriptions of several crossover designs, for example Zaph's BAMTM, that a 4th order acoustic slope was achieved with a 2nd order electrical construct.

My question is, is this a universal outcome, or is it dependent on a set of variables (and what are they)? And given this, what then would be the acoustic slope of a 4th order electrical construct, under the same circumstances? 6th order? 8th order? What would be the acoustic slope of the BAMTM crossover if it were actually constructed 4th order electrical? When does the acoustic slope actually equal the electrical, or does it ever?

Please help the crossover challenged!

The acoustical slope is nothing more than the vector sum of the driver's (measured) acoustical response and the response of the electrical network. With a bi-amped low-level crossover, there is no electrical interaction between the complex load of the driver and the crossover filter. With a high-level passive network, there IS an electrical interaction (which must be compensated for in the filter design, typically using computer software).

For both types of crossover, though, FILTER + DRIVER RESPONSE = ACOUSTICAL SLOPE.

Note the term VECTOR sum. This means the phase shift of the driver is ADDED to the phase shift of the crossover. So, if the crossover is rotating the phase by 90 degrees at the crossover frequency (typical of a 2nd-order network), and the driver is adding another 90 degrees, you get a net phase rotation of 180 degrees at the crossover frequency. This is why a real-world acoustical measurement is always needed to confirm you're getting what you planned for.

[sdclc126]

Thanks for the excellent explanations - happy to hear from two very well-versed authorities in the hobby!

I had completely neglected natural driver roll-off as part of the equation! I was pretty sure individual driver characteristics play a role though, as this of course goes for all components.

Bottom line is never assume and always measure!

Thanks again guys!

P.S. - Lynn - can't wait for that dipole design!

[coldplug]

Hello good people!

I could not find any forum rule about reopening old topics so I will post my question here because it is closely related to this topic.

I do very well understand the term acoustic slope that is explained here. But again, Zaph Audio site use some terminology that I need to clarify myself.

Zaph say, for example:

The crossover is an acoustic 4th order at 1950 Hz, and is nearly a symmetrical Linkwitz-Riley thanks to the woofer's shallow acoustic center.

What I do not understand completely is that Linkwitz-Riley type of filter. In its basics I do understand what will LR mean in electrical terms, that is how to calculate passive components values and how the response will look around xover frequency. But, how is term "LR4" applied to acoustic response?

How one will create "acoustic Linkwitz-Riles 4th order" response, in comparision to, say, "acoustic Butterworth 4th order" or "acoustic Bessel 4th order" filters? What do these filter types present in acoustic terms?

Thanks if anyone will be so kind to clarify this!

[richie00boy]

It's about the Q of the filter, which affects the roll-off around resonance, the phase response and impulse response.

[Pano]

Hi Coldplug, welcome to the forum and thanks for looking at the rules! It is OK on this forum to post in old threads, it keeps the dust from settling.

The acoustic slopes are just like the electrical ones. A 4th order L-R electrical or acoustic would look the same. Ditto Butterworth and others.
Getting there, tho, may take a combination of electrical and acoustic slopes.

[coldplug]

Thanks,

so, we are talking about filtered acoustic response of the driver, that is, about the final shape of it's curve on the SPL graph, am I right? So, if I'm correct, LR acoustic slopes will have smoother rolloffs in xover region than Butterworth slope, and that's it?

OK, then if I understood that correctly, now let me ask you about a real word example. Let say you have two drivers and you have built enclosure for them. Mounted them and measured their respective response without any filters.

Now, what will you do if you want to create exact Linkwitz-Riley, 4th order xover for them at some frequency?

Does this mean:

a) fire up a crossover simulator, load measured data, key-in xover frequency and choose LR 4th order filter? I don't think so, because software will do electrical slopes, not acoustical, if I'm correct?

or

b) choose any filter in crossover simulator that gives you closest approximation of response that visually looks like LR slope, and then change capacitor and inductor values and maybe add some notch filters, until simulated response visually looks most closely like LR slope, completely regardless of what component values you get

or

c) any other approach?


Thanks!

[Pano]

Originally Posted by coldplug

so, we are talking about filtered acoustic response of the driver, that is, about the final shape of it's curve on the SPL graph, am I right?

Yes, you are right.

Now, what will you do if you want to create exact Linkwitz-Riley, 4th order xover for them at some frequency?

Measure the drivers in the box then fire up Passive Crossover Designer (PCD). Import the impedance and FR files. Choose a slope as your target. Use the PCD to find the component values that give you the target slope.
Build prototype crossover from PCD values then measure and tweak to get what you want.
..

[coldplug]

Originally Posted by Pano

Measure the drivers in the box then fire up Passive Crossover Designer (PCD). Import the impedance and FR files. Choose a slope as your target. Use the PCD to find the component values that give you the target slope.
Build prototype crossover from PCD values then measure and tweak to get what you want.
..

So, does that mean that Passive Crossover Designer will give you component values needed to reach acoustic slopes? If that's true then it's in fact easier than I thought.

Thx!