Hello all. I'm not very experienced, but gathering ideas to try for my next project and have seen advice to try to avoid any crossover in the 300-3000 hz, high sensitivity region. But when I sit down to try to lay out a design that could meet that requirement, that one condition seems to virtually dictate all the choices. Like this: XO to tweeter at say 3.5k demands center to center spacing of about 2 inches to prevent a hideously tight lobe at XO freq. That in turn requires both a 0.75-1 inch tweeter (so sorry, nothing cool like an AMT, just stick to domes) AND a 3" max mid...and exactly one of it, as a 2.5 way would require a "forbidden" XO and two playing together would either move their acoustic center too far away or, in an MTM, also worsen the lobing that we can just barely tolerate as is. So we cram a 1" tweeter and 3" mid as close together as we can, and do what we want for bass, but only to 250 Hz or so. And we've accepted compromises on all fronts: directivity, power handling, distortion, max SPL, driver selection, creative choice, fun....or marketing wow factor, if you insist on objective metrics 😁. But my question is: do the people who give this advice know how constraining it is, and still think it worth it? Is avoiding an XO in that range really so magical that it is worth collapsing the universe of possibilities so severely? Or am I being too absolutist about directivity in the XO region, and what is gained is anyway thought to be worth a wasp-waisted XO at 3.5k, which after all is not your dog dying? Are there particular XO tricks that mitigate the tight spacing/lobing trade off? (Perhaps I should look again at the original MTM with 3rd order + phase shift XO....?) Does midrange XO phobia abate significantly if the XO is active DSP (the only way that I will do business, being myself soldering phobic).
I guess somewhere in there I started to suspect that the 'no midrange XO' fan just values a transparent midrange...subjectively, a midrange about which they have no nagging doubts...more than some other things, notably directivity. That's not crazy. But I still have the questions:
1. Is it really that great? best case? - and,
2. How often is it actually done? (I think rarely?)
Opinions or insights? Thanks folks. 😁 Make a strong case and I guess I'll have to try it!
I guess somewhere in there I started to suspect that the 'no midrange XO' fan just values a transparent midrange...subjectively, a midrange about which they have no nagging doubts...more than some other things, notably directivity. That's not crazy. But I still have the questions:
1. Is it really that great? best case? - and,
2. How often is it actually done? (I think rarely?)
Opinions or insights? Thanks folks. 😁 Make a strong case and I guess I'll have to try it!
Some of those that object to a crossover in that range may not have heard one that has been implemented extremely well, or with drivers that truly complement each other.
You hit the nail on the head with the compromises, everyone has their list of compromises they can live with and those that they can't. Provided none of the ones you can't live with are mutually exclusive, then you should be able to come up with something that is right for you.
Tony.
You hit the nail on the head with the compromises, everyone has their list of compromises they can live with and those that they can't. Provided none of the ones you can't live with are mutually exclusive, then you should be able to come up with something that is right for you.
Tony.
A lot of 2-ways do this, and I honestly cannot hear a benefit to this theorem. Joseph Audio is at least one famous brand that is happy to cross at 1.8kHz with steep crossovers and no issues at all.
I also call a little ******** on the C to C spacing. At normal listening locations it really is not an issue, plus, come on! With a reasonable midrange and wide-baffle tweeter this gets stupid.
There is one idea that gets you past all of this, using a woofer-assisted wide-band. Use a driver that can cover 400 to 20 kHz reasonably well and add a woofer.
Finally, have you seen the latest XSim? Bill is adding this kind of simulation to it now, so you can see exactly how a design will sound at different directions. It may really help you get past this.
Best,
E
I also call a little ******** on the C to C spacing. At normal listening locations it really is not an issue, plus, come on! With a reasonable midrange and wide-baffle tweeter this gets stupid.
There is one idea that gets you past all of this, using a woofer-assisted wide-band. Use a driver that can cover 400 to 20 kHz reasonably well and add a woofer.
Finally, have you seen the latest XSim? Bill is adding this kind of simulation to it now, so you can see exactly how a design will sound at different directions. It may really help you get past this.
Best,
E
Last edited:
imo ideally not crossing at 120 - 4500 , not very feasible, it all depends on the drivers and choices of integration, like eriksquires said if some drivers can tolerate steep crossover and xo well in 1.8khz why not?
xo at 120 is easy. xo at 3500 demands a special full range driver. If you use only a 3 inches you will be able to reach the target frequency at price of compromise in the bottom range, however this will reduce your building costs. A good 8' full range is costly. A good 3'' is affordable. Most 5" to 8" woofers cross well for 2 to 2.5, its the very minority that can play well higher and have good resolution and foundation under 1khz (most important) .
xo at 120 is easy. xo at 3500 demands a special full range driver. If you use only a 3 inches you will be able to reach the target frequency at price of compromise in the bottom range, however this will reduce your building costs. A good 8' full range is costly. A good 3'' is affordable. Most 5" to 8" woofers cross well for 2 to 2.5, its the very minority that can play well higher and have good resolution and foundation under 1khz (most important) .
The whole "don't cross between <insert frequency range here>" probably comes from people who have heard systems with bad power response because they are crossed too high. Either that or they have convinced themselves that they can hear phase distortion. For the low filter orders that passive speakers use, phase distortion is not audible and in a complete 2-way/3-way passive system the phase continues to wrap in between XO points anyway.
Most 3" aren't able to cross below about 300-400Hz LR4 at a high volume (>100dB/1m) without significant distortion. People also underestimate centre to centre spacing, it's not that much better than say a 6.5" or 7" woofer. A 7" woofer really needs to be crossed no higher than 2kHz and ideally below 1.5kHz. Using a 3" mid and a full size tweeter you should still aim to cross at 2.5kHz or below if you value having smooth power response.
Most 3" aren't able to cross below about 300-400Hz LR4 at a high volume (>100dB/1m) without significant distortion. People also underestimate centre to centre spacing, it's not that much better than say a 6.5" or 7" woofer. A 7" woofer really needs to be crossed no higher than 2kHz and ideally below 1.5kHz. Using a 3" mid and a full size tweeter you should still aim to cross at 2.5kHz or below if you value having smooth power response.
Last edited:
There are competing unrelated problems as to where is easier to cross so the numbers are only suggestions, and not everyone agrees. Of course it becomes irrelevant if you can cross well.
No difference, this is unrelated.
I built a good sounding three way speaker two years ago with crossover points at 150 Hz and 800 Hz. 12" woofer, 7" midrange and four 1" full-range speakers as tweeters. These crossover points assured that each driver was in it's piston range ( no cone break up ). I used a fourth order crossover at the 800 Hz to minimize the audibility of the crossover.
I just wonder how many threads like this have been answered in the past.
"no XO in 300 Hz - 3kHz" does not have to be very constraining, several ways to solve it. Big CD + horn is one way, then you have acoustic panels of various kinds/Electrostatic designs and "Fullrange"/Wideband unit is another. And then there is a myriad of variations in implementation under each of these.
If you want "point source" you can also go the Synergy or Coaxial route.
But there are many ways to make a crossover in the 300-3khz range work very well with more traditional 2-way or 3-way designs. It is all about implementation.
Edit:
gabdx: Pro units like the Fane 12-250TC or the Fane FC152 (I am using this 15" FR daily) are not that expensive and can actually provide most of what you need with some tweaks, no need for extra help from a woofer if you do not want to.
"no XO in 300 Hz - 3kHz" does not have to be very constraining, several ways to solve it. Big CD + horn is one way, then you have acoustic panels of various kinds/Electrostatic designs and "Fullrange"/Wideband unit is another. And then there is a myriad of variations in implementation under each of these.
If you want "point source" you can also go the Synergy or Coaxial route.
But there are many ways to make a crossover in the 300-3khz range work very well with more traditional 2-way or 3-way designs. It is all about implementation.
Edit:
gabdx: Pro units like the Fane 12-250TC or the Fane FC152 (I am using this 15" FR daily) are not that expensive and can actually provide most of what you need with some tweaks, no need for extra help from a woofer if you do not want to.
Last edited:
This advice harks back to the days where crossover design was a lot harder, usually due to the lack of understanding and easy access to design software.
People would build the majority of their designs without any measurements. The large lump of advice that went with the *keep out* sign between 300-3000kHz was always pick drivers with smooth, flat frequency responses etc.
The whole point of this was a damage limitation exercise. If you start out with flat drivers and put your xover points away from where the ear is most sensitive then you're far less likely to end up with something terrible. Or rather if your botch the xover then at least you've limited the damage to where it isn't as easy to hear and your intrinsically flat drivers are helping to keep things decent elsewhere.
People would build the majority of their designs without any measurements. The large lump of advice that went with the *keep out* sign between 300-3000kHz was always pick drivers with smooth, flat frequency responses etc.
The whole point of this was a damage limitation exercise. If you start out with flat drivers and put your xover points away from where the ear is most sensitive then you're far less likely to end up with something terrible. Or rather if your botch the xover then at least you've limited the damage to where it isn't as easy to hear and your intrinsically flat drivers are helping to keep things decent elsewhere.
Well the ideal speakers should have a dedicated driver for each frequency to prevent response dips from the suspension countering the cone motion and polar response.
When you 'hear' what XO does (in my case) you can compromise, however you need then a higher quality driver which can somewhat sound ok covering a wider range.
My current preferred speakers are closed box and 5' covers from 50hz to 4500hz third order on the tweeter, second order on the woofer.
Yesterday I was browsing speakers and some very big drivers from Fostex can cross in the + range despite their size, they are costly and I don't know how they sound.
FW405N, 2k
FW305, 3k
F200A, (...)
and Ellipticor 18WE/4542T00 8", 4k maybe.
When you 'hear' what XO does (in my case) you can compromise, however you need then a higher quality driver which can somewhat sound ok covering a wider range.
My current preferred speakers are closed box and 5' covers from 50hz to 4500hz third order on the tweeter, second order on the woofer.
Yesterday I was browsing speakers and some very big drivers from Fostex can cross in the + range despite their size, they are costly and I don't know how they sound.
FW405N, 2k
FW305, 3k
F200A, (...)
and Ellipticor 18WE/4542T00 8", 4k maybe.
Crossovers typically induce phase shifts, and the drivers being crossed typically have different dispersion patterns.
People are probably most sensitive to phase issues in the 300 to 1000 hz range. Below 300 hz, the room overwhelms. Above 1000 hz, the wavelengths are too short for phase to matter much.
In contrast, Geddes believes that constant directivity is most important above 1000 hz. There's probably some point where it starts not to matter again... maybe 8-10 kHz.
Thus if you have drivers that can be crossed above 1 kHz without inducing changes in directivity, then there's little problem. You'll never notice the phase induced by the crossover.
If your drivers have different dispersion characteristics, it's probably best to cross below 1000 Hz. You'll induce some phase issues where you're most sensitive, but that's probably preferable to the dispersion issues.
Ideally you don't have any crossovers in the 300-1000 Hz band, and you manage to keep directivity constant (or at least smooth) above 1000 Hz.
People are probably most sensitive to phase issues in the 300 to 1000 hz range. Below 300 hz, the room overwhelms. Above 1000 hz, the wavelengths are too short for phase to matter much.
In contrast, Geddes believes that constant directivity is most important above 1000 hz. There's probably some point where it starts not to matter again... maybe 8-10 kHz.
Thus if you have drivers that can be crossed above 1 kHz without inducing changes in directivity, then there's little problem. You'll never notice the phase induced by the crossover.
If your drivers have different dispersion characteristics, it's probably best to cross below 1000 Hz. You'll induce some phase issues where you're most sensitive, but that's probably preferable to the dispersion issues.
Ideally you don't have any crossovers in the 300-1000 Hz band, and you manage to keep directivity constant (or at least smooth) above 1000 Hz.
Hi olsond3, can you tell me more about the top end in your previous design? Was tweeter layout 2x2, 1x4, other? Pointed same direction? I was just looking at the idea of multiple tweeters pointed different directions on a cylindrical baffle, and turning on the side tweeters to support HF dispersion as the main T starts to beam. (Concluded: seems like too much trouble for too little benefit.) But I would not have thought to cross to a 1" driver at 800 hz, that's creative. Did you trouble about directivity around that M-T xo (mitigated by layout?) or just found it sounded good?
Ah, thanks for the pointer to XSim. I have been building rather than thinking about design for a while....got to check out the CAD tools!
It's true I was dismissing a number of options without discussion. I'm familiar with most of the approaches you named. But what is 'synergy'?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.