Is there an optimal number of crossover points that a speaker system should have in a multi-way system?
I would imagine that the smaller the frequency range that a driver has to produce, the better it would perform at those frequencies which leads me to wonder if more = better.
But I know there must be more to it?
I would imagine that the smaller the frequency range that a driver has to produce, the better it would perform at those frequencies which leads me to wonder if more = better.
But I know there must be more to it?
Direct radiating, or horn loaded? The answer changes because of the great differences in modulation distortion between these two types. The modulation side bands levels are usually 20-25 dB lower for horns using the same drivers as in direct radiating case at the same on-axis measurement. This is pretty big deal as you increase the SPL past 80 dB on-axis one metre.
For direct radiating, Dunlavy used 4-way first-order crossovers, and apparently achieve spectacular results (...for a direct radiating loudspeaker).
Chris
For direct radiating, Dunlavy used 4-way first-order crossovers, and apparently achieve spectacular results (...for a direct radiating loudspeaker).
Chris
From my own personal experience, 4-way seems best. Here are the reasons. When you have two way, there are compromises being made, like when you select big woofer, you have to cross low, that puts big demand on the tweeter, which have to work hard. You could use small fulrange for tweeter duties, but that is another story. So you move to 3way. That solves one issue, but now you have two crossover points, typically 300, 3k, and it may work great. But people often like to use ribbons, which needs to be crossed high, so you introduce upper mid. That way you can have really great subwoofer, covering say 20 to 150, then nice midbass taking over from 150 to say 800, then nice upper midrange, up to 5k, from there to ribbon. If all the drivers behave well, you can end up with great sounding speaker. I made few like that, with hivi planar tweeter, two 2" domes as mtm, two 6" midbases, and 15" woofers. Typically I biamp all my systems. Active crossover (Hiraga at 100Hz), classD for woofer, classA for rest.
Cheers.
Cheers.
It has to do with the drivers (and radiating baffles/horns), not the crossover networks--ideal or not. Crossover networks bring their own set of problems if you are trying to use anything but first order filters: time misalignment, group delay growth, and phase errors.
In general, the fewer the crossovers, the better. But this is always traded against what the drivers can do and their distortion, particularly as the on-axis SPL rises above 80 dBC/1 m.
I personally recommend two-way full-range horn loaded multiple entry horns (MEHs) as the best choice among alternatives in terms of fidelity and directivity that you can make. But you can also get dual-diaphragm 2" compression drivers which break the higher frequency band into two parts in order to minimize FM distortion--which is a function of the bandwidth of the driver and the excursion of the driver diaphragm. That makes the design a three-way and is not limited (effectively...for home use) on the higher end of SPL like direct radiating configurations.
If you're using horn-loaded drivers, their diaphragms only have to move about 1/5th the distance as the same driver being used in direct radiating mode...to achieve the same on-axis SPL.
So there isn't a pat answer to your question (as you probably already could guess).
Chris
In general, the fewer the crossovers, the better. But this is always traded against what the drivers can do and their distortion, particularly as the on-axis SPL rises above 80 dBC/1 m.
I personally recommend two-way full-range horn loaded multiple entry horns (MEHs) as the best choice among alternatives in terms of fidelity and directivity that you can make. But you can also get dual-diaphragm 2" compression drivers which break the higher frequency band into two parts in order to minimize FM distortion--which is a function of the bandwidth of the driver and the excursion of the driver diaphragm. That makes the design a three-way and is not limited (effectively...for home use) on the higher end of SPL like direct radiating configurations.
If you're using horn-loaded drivers, their diaphragms only have to move about 1/5th the distance as the same driver being used in direct radiating mode...to achieve the same on-axis SPL.
So there isn't a pat answer to your question (as you probably already could guess).
Chris
Last edited:
I think it's difficult to pull off a good 2-way system. Sadly this is what is churned out by many loudspeaker DIYers. So for boxed, direct radiator type loudspeakers, IMHO 3-way is the minimum for covering nearly the full audio spectrum without too many deficiencies. It's also more costly, and more complicated, to make a 3-way passive crossover.
For open baffle and dipole systems, the requirement gets bumped up to a 4-way system due to "low frequency" dipole cancellation. It's best to handle the crossover duties with some kind of active crossover system (including DSP).
The addition of a subwoofer dedicated to reproducing the lowest octave(s) can still be added on to these systems and will relieve the woofer's excursion requirements and reduce the amount of distortion for that driver.
For open baffle and dipole systems, the requirement gets bumped up to a 4-way system due to "low frequency" dipole cancellation. It's best to handle the crossover duties with some kind of active crossover system (including DSP).
The addition of a subwoofer dedicated to reproducing the lowest octave(s) can still be added on to these systems and will relieve the woofer's excursion requirements and reduce the amount of distortion for that driver.
Last edited:
I think it just comes down to how loud do you need to get?
(Where loud includes excellent quality)
IMO, we go multi-way because it's the only way to raise SPL equally across all 10 octaves.
As a rule, the louder a driver needs to run, the narrower its designed bandwidth.
Which of course equals more bands, more crossovers.
I am quite satisfied with 3-way in my open baffles.
PA woofers suitable for open baffles can be crossed high, it helps much.
Another justification for 3 way is DSP crossover, with FIR cutoff filters.
FIR filters very effectively protect the speakers from overdriving, you have a slightly increased freedom to choose the crossing points.
Additionally one can make different presets, with different tradeoffs.
PA woofers suitable for open baffles can be crossed high, it helps much.
Another justification for 3 way is DSP crossover, with FIR cutoff filters.
FIR filters very effectively protect the speakers from overdriving, you have a slightly increased freedom to choose the crossing points.
Additionally one can make different presets, with different tradeoffs.
The theoretical answer is obvious: a single driver is best.
But only a large electrostatic panel can nicely play enough of the range (say, down to 140 Hz to 14 kHz).
Next best would be a driver which has no crossover, is seamless, and covers the whole middle range, leaving specialized units to handle just an octave or two at the top and bottom. But whenever you have a crossover that low, you have to go to DSP and bi-amping.
B.
But only a large electrostatic panel can nicely play enough of the range (say, down to 140 Hz to 14 kHz).
Next best would be a driver which has no crossover, is seamless, and covers the whole middle range, leaving specialized units to handle just an octave or two at the top and bottom. But whenever you have a crossover that low, you have to go to DSP and bi-amping.
B.
4-way with dsp-xo gives you best chances to achieve high fidelty.
Define what you are trying to achieve (max spl, bass extension, bass alignment, radiation pattern/directivity, distortion). Choose drivers, measure and study each way's response also offaxis and off-band, (optionally simulate), equalize, find best xo type for each x, set delays and spls. Then reiterate... Simple as that!
Define what you are trying to achieve (max spl, bass extension, bass alignment, radiation pattern/directivity, distortion). Choose drivers, measure and study each way's response also offaxis and off-band, (optionally simulate), equalize, find best xo type for each x, set delays and spls. Then reiterate... Simple as that!
Last edited:
So that includes harmonic, phase/group delay/time alignment, modulation, compression distortions.
Did you think about power response/polar coverage in your definition of "distortion"? It happens to be the most important type of distortion in terms of subjective listening, as reported by Sean Olive at JBL/Harman.
Chris
Did you think about power response/polar coverage in your definition of "distortion"? It happens to be the most important type of distortion in terms of subjective listening, as reported by Sean Olive at JBL/Harman.
Chris
The more crossover points, especially when slopes are steep, the worse overall problem with phase shift/group delay.
But generally, hearing of many people seems to be forgiving in this respect.
If you've listened to multi-way systems with IIR filters and you liked the sound, then you have a lot of crossover solutions to choose from.
Yep, one, obviously, between your woofer and wideband driver as low as you can possibly get away with. Even better if you do away with the cabinet too
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.