advanatge of 1s order over 2nd or 3rd order?

Curmudgeon

Member
2005-11-02 7:13 pm
First, I'd refer you to Vance Dickason's most excellent Cookbook.

First order crossovers with almost any driver will require significant care in handling the tradeoffs. If nothing else, it tends to require (expensive) drivers with excellent bandwidth and power handling.

From my own experience, pure first order, that is acoustic first order response in and far out of band really is not feasible. The multi-slope first order, popularized by Thiel uses a first order roll-off at the crossover frequency, followed by another pole about an octave or so out, by which time the driver's rolloffs kick in.

High end breakup on the woofer and mid usually need to be trapped out, as the breakup is likely to be only 15-25 dB down, which, even if not measurable on a composite frequency response curve, will be audible.

If the mid/tweet voice coils are vertically aligned (or more accurately, the sound sources) there will be a -15 degree slope of the smoothest axis. Hence the tilt back of some speaker systems.

Aside from the (frequently disputed) phase linearity issues, I like it because it provides a broad crossover region which can often sound seamless.

I squuz in this reply before leaving for dinner appt., so it's probably even less coherent than usual. Apologies.
 

Curmudgeon

Member
2005-11-02 7:13 pm
I am speaking of 1st order crossover network although I am also interested in the acoustic slope too.

Curmudgdeon you say But how does this affect the off axis response?

The crossover behavior is based on the combined electrical and acoustic responses; so for a true first order, that would be one pole electrical rolloff (-6 dB/octave) and NO acoustic rolloff.

I think you're asking about my use of the word "broad"? I meant broad frequency span; the overlap is roughly an octave.
 
First, I'd refer you to Vance Dickason's most excellent Cookbook.

If the mid/tweet voice coils are vertically aligned (or more accurately, the sound sources) there will be a -15 degree slope of the smoothest axis. Hence the tilt back of some speaker systems. .


That should read "If the mid/tweet voice coils are NOT vertically aligned (or more accurately, the sound sources) there will be a -15 degree slope of the smoothest axis. Hence the tilt back of some speaker systems."

I refer you to Vance Dickason's most excellent Cookbook. :)
 

Curmudgeon

Member
2005-11-02 7:13 pm
Loudspeaker Design Cookbook, by Vance Dickason, 7th edn, P. 157. "Because of the 90 degree phase difference, a -15 degree tilt will occur in the vertical polar response with high-pass and low-pass driver separated by a distance of one wavelength at the crossover frequency (+15 degrees for the reversed polarity connection)." Acoustic centers are assumed aligned for obvious reasons.

Figure 7.8 shows the negative max output lobe, the dotty line is the normal polarity, and the heavy line is the reverse polarity. Figure 7.10D shows the tilted cabinet for a horizontal max output lobe.

It's the even order alignments that do not have the tilt.
 
Last edited:
In an electrical engineering class, a first order highpass plus first order lowpass sum more perfectly than steeper slopes.

In real life with real drivers, it just doesn't work that way because the drivers have some inherent phase and magnitude changes versus frequency.

If you read Vance's book enough, and read other posts in threads with "crossover" in the title, what you'll find is that you have to really measure the drivers and make a kind of empirical design. All that theoretical stuff is great as a background to understand what is going on, and maybe useful for electronic crossovers, but doesn't apply well with real drivers.
 
In an electrical engineering class, a first order highpass plus first order lowpass sum more perfectly than steeper slopes.

In real life with real drivers, it just doesn't work that way because the drivers have some inherent phase and magnitude changes versus frequency.

If you read Vance's book enough, and read other posts in threads with "crossover" in the title, what you'll find is that you have to really measure the drivers and make a kind of empirical design. All that theoretical stuff is great as a background to understand what is going on, and maybe useful for electronic crossovers, but doesn't apply well with real drivers.

I completely agree. Having said that, first order design is generally a lot easier. What many get wrong is driver selection. This IMHO is the single most important part of loudspeaker design. Get this right and everything falls into place. Get it wrong and you spend lots of time and lots of money making the crossover tame the drivers.

Terry
 
Last edited:
I completely agree. Having said that, first order design is generally a lot easier. What many get wrong is driver selection. This IMHO is the single most important part of loudspeaker design. Get this right and everything falls into place. Get it wrong and you spend lots of time and lots of money making the crossover tame the drivers.

Terry

True dat !!!!! ...........................:drink:
 
Hi Mate, I have the fifth edition but let us see if we can agree on this. From what you have said the answer in your book may be in 7.10. Compare A,B,C, and D. The only diagram with acoustic centres on the same plane is C. This is why most quality speakers with 1st order slopes use a stepped baffle.

Terry
Hmm, I was a tad hasty on the figure, which merely demonstrates two cases with the acoustic centers non-aligned, and two (not one) aligned, one with a stepped baffle, and one with a tilted baffle.
However, that is not the basis for my disagreement, which is that you contradicted my post which stated that for a first order crossover, normal polarity, acoustic centers aligned, there will be a -15 degree negative slope to the main lobe. See p149, (the discussion of figure 7.8) which repeats this. It's also a design I've used, and am very familiar with.

Construction methods used to align the acoustic centers are really irrelevant to that.

You certainly have a point head_unit; first order, more than any other alignment, has stringent driver requirements to reasonably satisfy the criteria. And, of course, while there are no perfect drivers, there are some very good ones...

In the late '80's I used Scanspeak 8" polyprop woofer, the little 8636? mid, and the big, vented, Alnico Focal tweeter; each had the spectral and dynamic range to be usable in a multi-slope. A more recent design, for personal use, used the Scanspeak 25W pulp/carbon woofer, Skaaning Sk4-H52, and the Scanspeak 9000 Revelator tweeter. Notches (traps) were needed for the out of band breakup of the mid and woofer.

That's why I think multislope is necessary, as a really good pure first order system would be a heroic undertaking, and given driver acoustic rolloffs, probably impossible.
 
Last edited:
Curmudgeon we will have to agree to disagree then. If you can put up the four figures in a post I can explain further. Figures A and B have the drivers in the same plane while figure C has the Acoustic centres in the same plane. Figure D is one solution where a sloping baffle brings the drivers in line with the acoustic centres. It has it's draw backs. The primary one being that you are forced to listen at 15 degrees off axis.

If you choose the right drivers first order can be a breeze. For those that may wish to: A morel MW168 or MW 144 with a Morel DMS 37 tweeter is a great combo. Surface mount the woofer and rebate the tweeter on a flat baffle. The acoustic centres are aligned! The crossover has two components only. Yes, the drivers are expensive but the x-over is not!

Terry
 
I go with the lowest order that gets the job done. I have a preference for 1st order (electrical) series but that requires careful selection of drivers which does take a lot of time. Get that right and you will be rewarded. I've only been able to do it on less than half of my speakers and the others are generally AR series which use higher orders.... 2nd to 3rd order acoustic.

I've noticed a lot of manufacturers in recent times have being using lower orders and also noticed some low order tests by Troels which may indicate a low order crossover speaker on the way from him.
 

Caleb Chia

Member
2008-05-20 11:17 am
How about phase coherent active crossovers as described in Randy Slones' book The Audiophiles Project Sourcebook, I think it was. He said they worked by subtracting the unwanted frequencies from the signal and so had a even phase shift for all frequencies or something. Would this be the optimum?
The circuits looked quite complex on the higher order xovers though.
 
Was it dantheman who linked to a little applet that showed phase interference. Am not seeing that one right now maybe someone can link to that again.

Lower order crossovers have less stored energy. Higher order crossovers with low Q store less energy than their high Q brothers. Stored energy and what it does to transient response might be important.

This link (((acourate)))® - Room Correction, Speaker Optimization and Sound Improvement leads to software for Linux computers which operate sound cards using FIR type filters which are both very steep and lack phase shift in the pass band. Have heard nothing but great things about this. :)