does lower order xover have any advantages?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am new to speaker design and currently planning on building a pair of speakers. It seems like the xover is the most difficult part of the project.

My question is does 2nd order have any advantages over 4th order xovers?

I means with 4th order you can get a more flat frequency response but with all kinds of components in the xover, how will the sound get affected?

For example I heard that notch filter can eliminate certain frequency anomalies but can affect the transient response and transparency.

What would be a compromise?

What would be a pair of woofer/tweeter that should work well with a 2nd order xover?
 
diyAudio Moderator Emeritus
Joined 2001
The lower the order, the less the phase shift. Also, the less the tendency to "ring".

However, also the lower the order, the greater is the number of octaves of overlap where both drivers are contributing significantly to the hearable output.

Up through the 1970's, most speakers had either first order or second order designs in their crossovers. Many still do, but now many go for higher order designs in their crossover as well.

The advantages of the second order over the fourth order are, therefore, fewer components, less tendency to "ring".

The advantages of the 4th order over the second order are: fewer octaves of overlap between drivers, and 4th order does NOT require one driver be hooked up out-of-phase to the other, as second order does.
 
Get a copy of Vance Dickason's Loudspeaker Cookbook, it covers the details of all these things to the point where your head will hurt trying to digest it all... but a very useful thing to have around.

First order filters will sum properly and result in an impulse waveform that is quite nice... nothing else will short of no xover.

2nd order filters are a problem because the phase of the drivers in the pass band is flipped 180 degrees (out of phase) by the filter. So, you can flip the polarity of the driver, but then the phase in the xover region is not right.

2nd & 3rd order filter don't sum properly in the passband, IF you manage to get them to act like they should in that region.

4th order filters flip the phase 360 degrees - in phase but time delayed, and have the benefit of "getting the energy out" of the stop band quickly (useful for running a driver right to the hairy end of its usable response). They don't sum right either in the pass band.

Linkwitz figured out if he dorked the filter's Q and moved the nominal xover points a bit he could get the thing in 4th order to sum pretty nicely. That's a Linkwitz.

In practice unless your drivers are FLAT and without excessive phase shift at least + & - a full octave on either side of the xover point, your *electrical* values that look good on paper or in the computer simulation will not give you the expected slopes or roll offs at all.

Since few drivers have these characteristics, few xovers work the way they are planned. What you end up doing is empirically adjusting the design to best fit the desired result. The more you know about the effects and the limitations of the driver and the electrical circuit design, the better shot you have at getting a really good result.

As is obvious from the cheap-o commercial speakers, anyone can and does plunk a xover into a box with speakers and get sound out that appears fairly reasonable.

Start with what you can do - measure with some computer aided stuff and a microphone, and dork it, go through the learning curve. There's no easy route other than building a clone of a known good design with the exact and identical componenets and parts.

Changing the parts and the box changes the result.

Enjoy!

Welcome to the world of loudspeakers, design and infinite compromises and variations! :D

_-_-bear :Pawprint:
 
"First order filters will sum properly and result in an impulse waveform that is quite nice... nothing else will short of no xover."

Yes, there are other alternatives. You familiar with JohnK´s work?

"2nd & 3rd order filter don't sum properly in the passband, IF you manage to get them to act like they should in that region."

2nd LR sums flat as does 3rd order BW and 4th order LR. 2nd order BW and 4th order Bessel does for example, does not sum flat though.

This according to the reference you mention yourself. :)

/Peter
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.