Strange idea.... (1 Viewer)

[Gianluca]

I just got an idea...

What if I built a 5 ways speaker system?

1" tweeter;
3" midrange;
6" midrange;
8" woofer;
12"(sub)woofer.

each of the drivers would be timbre matching and all of them in a different box, with external Xovers.

Would this configuration have any results?

This is not a joke... I'd like to hear some opinions from the experienced DIYers.

Thanks.

 

[TimForman]

Yeah, come you guys. Give us the crossover design!

 

[Anthony_Gomez]

2 ways are challenging
3 ways aren't for the faint of heart.
4 ways...have only seen done successfully by some VERY experienced people ...rarely.
5 ways ....ummm...no thank you!!!

:-D

 

[Allen Ross]

what about active crossing?

 

[ThomasW]

This is not a joke... I'd like to hear some opinions from the experienced DIYers.

Well this VERY experienced builder suggests using the fewest number of drivers and XO points for the best performance.

A good two way combined with a sub offers the best performance IMHO.......

 

[JesseSilver]

I'm not an expert, by any means, but I know what sounds good to me. My speakers are a pair of Dunlavy SM-1's and a McCormack amp. Those speakers (which are just tweeters and 6" mids) are only off -1.5db at 60Hz. Many other brands and DIY speakers are capable of similar performance, so I'm not quite sure what you'd gain by adding more and more drivers other than coloration from the crossover points. Sure, you might get better dynamics, but how loud do we all listen anyway?! Personally, I LOVE my speakers and once I finish my AV15 subwoofer to fill in the area below 55, I think I'll be one happy clam

Like I said, I'm not an expert... that's just my opinion. Less is almost always more.

 

[RichardHOS]

IMO, there's two ways to go in pursuing the "ultimate" speaker. One is to keep crossover points out of the mid-treble region where vocals and critical instruments are voiced, from roughly 300Hz to 3000Hz. That would result in a three way speaker (four if you need or count a dedicated sub for the bottom octave or so). The other approach is to use all active crossovers, and from the beginning concede that there will be crossovers in that critical region. Once you have one crossover point there, you might as well have several. Using active crossovers with steep slopes, you can confine the used bandwidth of a particular driver to a small portion of its linear range; the narrow used bandwidth and steep slope should give well behaved phase response. This could result in a five way, six way, or even more complex design. With phase problems out of the way, crossover design would actually be simpler than the typical three-way. Physical mounting of the drivers to avoid comb filtering artifacts is probably the biggest hurdle, and perhaps why I haven't seen any serious attempts at this design. It would seem to be well suited to a semi-line array type of speaker, with frequency tapering instead of power tapering. I think it's been done crudely, but I've never seen a full-out attempt at it.

 

[Hank Frankenberg]

What if you wanted clean, accurate sound instead?

Heed ThomasW's advice. He is spot on. You could spend a year learning how to design an excellent passive crossover for two-ways and maybe three-ways, but beyond that, uh-uh.
I'm not joking either - just trying to save you time and frustration.

 

[Bill Fagal]

Let me first say that I'm a proponent of full/wide-range drivers with a minimum of XO points. However, I can respect the potential benefits of multi-banding.

My thinking on this matter is similar to Richard's, with the addition that I would suggest digital crossovers for an all-out attempt to preserve phase relationships through steep XO points.

The other option no one has mentioned yet is multiple wide-range drivers crossed over fist-order. If you choose drivers with, say, six good octaves of bandwidth (and controlled out-of-band behavior) and limit each driver to its 2 middle octaves (or less) with 1st-order crossovers, you might end up with something nice, though you'd have to pay attention to lobing. Driver selection would be critical here, and suitable drivers may not be available.

Bill

 

[John E Janowitz]

The problem with a type of system like this is getting everything phase matched properly. In a two way system you design for flat response, then tweak component values until your phase is coherent. In a 2 way you can play with all the values in the mid and tweeter xover to get your phase correct. Changing each component value will shift your phase one way or another. So if you do a 3rd order on mid and tweeter, you have 6 components you can play with to adjust the phase.

In a 3 way or system with even more drivers, this becomes difficult. Say you get your xover phase set well between the woofer and mid. Now you move on to the tweeter. The only thing you can adjust easily without messing up the woofer/mid phase are the tweeter values. If you adjust something in the mid xover, it is possible to throw off the phase for the mid/woofer xover. So you only have 3 components in the tweeter section to play with to adjust the phase instead of 6 like in a 2 way.

This is the challenge I see with more drivers in a system. If you have drivers with a very flat response, and go through and do impedance compensation before you start the xover design it does become easier though. If everything had a flat impedance and flat response to begin with, Xover design would be simple.

John

 

[Jack Gilvey]

each of the drivers would be timbre matching and all of them in a different box, with external Xovers

How would you timbre-match the (sub)woofer and the tweeter?

 

[Anthony_Gomez]

How would you timbre-match the (sub)woofer and the tweeter?

easy....ditch the idea of a silk or metal dome tweeter and instead use a wood pulp tweeter. just make sure the tweeter and the woofer are made using neighboring parts of the same tree to get the best "timber" matching

 

[Hank Frankenberg]

deleted:b

 

[Hank Frankenberg]

Arrrggg...Tony, no...groan...:rolleyes

 

[RichardHOS]

Bill, the possibility of using first order XO's had crossed my mind, but I figured finding enough drivers with the requisite bandwidth, center frequency, and other quality attributes would be difficult. Perhaps a series of custom drivers would be in order? Why not... if someone wants to tackle a four, five, or six-way system, what's designing a couple of drivers along the way. Actually, seeing that Vandersteen and a few others do/have used first order crossovers, there are probably some drivers that could at least get you close. I guess the lobing wouldn't be any worse than your run-of-the-mill line array, and in reality should be better with the frequency tapering. I've thought about this before, and I think that done correctly this is probably the ultimate "no holds barred" approach (there are several "high-end" designs that follow this general scheme, though they use varying XO slopes and I don't think they've put all the pieces together).

And yeah - digital crossovers would be nice. Though, if I understand things properly, you can create arbitrary miminum phase crossover slopes in the digital realm but with an impact on transient response (i.e., you might get ringing). Wouldn't it be nice if CD/DV/SACD players sent digital information to a processor where all EQ, processing, and crossover functions were handled, with a suite of DAC's after that? Sounds like a Meridian system I suppose.


John, I don't think phase would actually be that much of a nightmare. First, by using active crossovers (or digital as Bill suggested) you can go to 24dB L-R designs without the complications of large capacitor and/or inductor values. Second, but using only a narrow segment of any particular drivers linear bandwidth, any abberant phase response of the speaker will be well down in dB. Some minor tweaking might be in order, but a "brute force" approach would probably be better than most would guess.

lol... I'm not going to spend my money to try it though.

 

[Bill Fagal]

Wouldn't it be nice if CD/DV/SACD players sent digital information to a processor where all EQ, processing, and crossover functions were handled, with a suite of DAC's after that? Sounds like a Meridian system I suppose.

Actually, that's my understanding of a Behringer DCX2496: AES/EBU in from digital source > digital 6-way crossover (6-48dB/oct.) > assignable digital parametric EQs > digital delays for each channel > 6 DACs

At $350, I am sorely tempted (and will likely succumb).

 

[Anthony_Gomez]

Hank....whaaaaat?

 

[RichardHOS]

Man, that would be about perfect if pre/pro's like the Lex MC8 had digital outputs for each channel so that you could use that Behringer (or something like it) in combination with DD/DTS/L7/etc. processing.

 

[John E Janowitz]

Richard,

Phase is the biggest nightmare of all. Take a woofer and tweeter, mount them on a flat baffle. The woofer's center is set back from the tweeter. Your phase is already out of wack. Just doing an electronic xover won't fix this problem. You'll still need a way to adjust phase independent of frequency response.

Getting response flat is pretty easy. Just about anyone can do that. Getting response flat and phase correct at the same time is a big challenge.

John

 

[RichardHOS]

IMO, "phase coherent" is over-rated and over-stated. Even using first order XO's and quality drivers, the impedance curves will deny you being phase coherent at any frequency other than the XO point. And if you achieve that, and other things are as they should be, you get a flat amplitude response through the crossover region. Funny thing is that with most common XO designs (2nd order, 3rd order) used in passive networks, being "in phase" between two drivers outside the crossover point is not only practically impossible, but undesireable, as it would cause a spike in the amplitude response.

I'm not a master speaker builder by any means, but I know enough to realize that using phase coherent 24dB L-R filters will go a long way towards elimnating and phase nasties you might encounter in such a system.

Again, I don't think it will be a walk in the park, but a fully active five-way wouldn't be nearly the nightmare as a passive five-way.