Which XO to choose?

[Whalebiter]

I'm designing new filters for my current project which at the moment has a 3rd order butterworth filter at 400Hz. The basses are 2 x Seas CA18RNX H1215 in 34ltr ported and the midrange is Focal 5K413S in 2.5ltr closed.

I thought I'd try a lower x-over, maybe at around 200 and I also want to use a 2nd order to get the serial resistance as low as possible and at the same time steep enough.

Problem. I can't seem to decide between Butterworth or Linkwitz Riley. I could of course just try both but because I plan to use rather expensive transformer inductors I thought I'd give this forum a try. How will the different filters sound and which in your opinion will fit best?

Btw in the butterworth case I plan to "stretch" the x-over by a factor of 1.3 to get to -6db at the x-over instead of -3db as in the original formula, so a 200 Hz x-over would give a 154Hz LP and 260Hx HP.

Any input would be greatly appritiated.

[PeteMcK]

my choice would be first order, bi-amped, so you can deal with Baffle step by level adjustment of the woofer.
If you're staying passive, I'd have the xover freq at the baffle step freq with same order xovers on both drivers, if 2nd order I'd use LR for flattest response

[tinitus]

Originally Posted by Whalebiter

Btw in the butterworth case I plan to "stretch" the x-over by a factor of 1.3 to get to -6db at the x-over instead of -3db as in the original formula, so a 200 Hz x-over would give a 154Hz LP and 260Hx HP.

You cant calculate like that, so precisely
It will only be estimated
If you dont use measurements you have your ears to decide
And approaching -6db at xo points is more like LR and not Butterworth
Straight electrical 12db will probably be anything but 12db, acoustical
But 12db as a goal is fine
Some kind of LR/Bessel mix might do fine
But component values may be different from what you expect
If you use a calculator you have to fiddle with impedance
But be prepared to make adjustments
So dont go buy the expencive stuff right away

As Pete says, you need to fiddle a bit to get baffelstep BSC right
But with double woofers you might end with crossing them low, and shallow, like 6db with a zobel
The rest might be more like pure 12db
Tho, midrange lowpass might be like woofers, 6db with zobel
You could try with resistors in series with paralel inductors
I have had good responses on that, but it can be tricky to get right

If you get a toroid(picture) for woofers, you can experiment with additional winding for a second inductor
Just be aware that value/induction changes accordingly with how you connect it
You need to measure induction to be certain
But it will be adviceable to measure and adjust all inductors anyway
And buy a collection a small cheap inductors for experiments
And be prepared that changes in one place of your xo may cause changes in other parts
For easy adjustmnts stay away from fancy connections, but do th layout exactly as in shown schematic
And use simple solder connection, at least until you are certain its finished, after months of listening
Remember that a lot of it is about phase
In the end, phase is what you need to look you fore
Good luck

[Whalebiter]

Thanks for the input guys but baffle step really isn't a problem. It might be because of my listening room, my equipment but i suspect it is the powerful Seas basses. That is in fact the reason I wanted to lower the xo in the first place, to get it so low that I can regulate (lower) mid and upper bass with the inductor. That's also the reason I wanted to go from 3rd to 2nd order - to lower DCR and get better control.

The LR filter and the BW filters have different roll off characteristics even though they're both 12db. The LR starts the rolloff earlier than the BW giving it another q-value. My question was simply how it will affect the overall sound at the xo area.

[LineArray]

Maybe the question here is whether to design for constant
sound pressure (LR -6 dB voltage drop) at crossover or
constant power (BW-3dB voltage drop).

My general impression is that constant sound pressure is preferable as
a tendency, when one or both drivers are already directional at
XO frequency. Otherwise the crossover region may be overpronounced.
This is for comparison of LR with BW of even order.

A BW of uneven order can match both conditions.


In the case of both drivers to be crossed beeing nondirectional,
optimization for constant power is preferable, which would be the
case for a typical subwoofer-satellite situation.

This is meant as the effective slope including natural rolloff of the drivers.

To achieve that, the slopes chosen for high- and low pass may look
rather asymmetric and of course room and positioning issues play a role
too. So it is a typical "it depends" kind of thing.

A large damped room may go to "constant sound pressure" a small
underdamped room may go to "constant power".

Whether the difference between your "stretched" (spreaded?)
BW and regular RL at order 2 is audible ... might be an academic question.

[Whalebiter]

Originally Posted by LineArray

Whether the difference between your "stretched" (spreaded?)
BW and regular RL at order 2 is audible ... might be an academic question.

Thanks yes I was beginning to think I wasn't able to get my point across. My technical english isn't the best. Let me show what I mean with some images from The Loudspeaker Cookbook 5th Edition.

Figure 7.25 is a normal BW
Figure 7.27 is a spreaded BW (Factor 1.3)
Figure 7.29 is a LR

As you can see the characteristics are different even though the outcome is almost the same. The parts used for the different filters are very different as well. My question was, as LineArray hinted to as an academic question, how they will differ other than output. I.e. are there differences in dynamics, bass control and so on.

[LineArray]

Hi,

i do not know either if "spreaded" is the correct term ...
at least we seem to know what we are talking about.

I expect the LR version to have phase shifting with
frequency more smoothly and have a more beautiful
impulse response.

Whether this is audible in a blind test - i don't know.

[LineArray]

For 4th Order i looked up a graph showing step response.
Lowpass BW reaches 100% somewhat earlier, overshoot
is around 20% for both (little less for BW) but
LR shows aperiodic behaviour ... while BW falls under
100% again after overshoot before stabilizing.

For 2nd Order the differences should be less significant.

Since your lowpass is tuned lower than nominal, you will
not reach 100% earlier with BW.

LR all in all has better step response than the
"spreaded BW" variant.

[Whalebiter]

Originally Posted by LineArray

So all in all LR has better step response.

Thanks then it's settled, I'll start off with LR and take it from there.

A picture of the project without finish, still very much in development. The tweeter is Scan Speak 9800 which has a xo @ 3k 4th order LR, which I'm also planning to lower.

[LineArray]

Has the LR lowpass larger inductivity ?

Then DC Resistance of the coil is somewhat higher.

Would you like to post the values for both alternatives ?