About crossover

Generally, the simpler the better.

First order series XOs seems to be the mantra of the day.

Now one thing to keep in mind, though. The lower the order of the XO, the higher the demands on the drivers are. Both on having a linear frequency response outside their primary working area and being able to cope with more signal there as well.
 
The best XO is the one that works best with the drivers you are using.

It is really hard to discuss XOs outside of the context of the speakers they are being used with.

The Full Rangers would say no XO is the best XO.

Personally, i am using a 4th order active XO between my woofers & full-ranges (@125 Hz), the FRs are allowed to run all the way up, and then a single cap XO on the superT at about 10k.

A 1st order XO allows you to get the response to sum to 1 in both the amplitude & the phase plane. But it puts big demands on the drivers you are using and to achieve all the promise the acoustic centres of the drives should coincide (at least in the vertical plane).

Faster XOs mean that you can run a driver closer to its edges, and you have less driver interaction. The Joseph speakers are an example of taking high order XOs to the max -- and they seem to be very highly regarded.

dave
 

argo

Member
2001-02-12 3:35 pm
Estonia
billy

Second both answers:
The best XO is one that works best with the drivers you are using - that is, if the drivers will demand even 6th or 8th order to sum flat and keep all resonant breakups at low level enough, this is the way to go.
But also try to keep the circuit as simple as possible - targeting a total response of +-3dB with resonant breakups lower than - 40dB is quite acceptable in most cases.
Sometimes you can do away without using all these resonant trap, impedance compensating and baffle step compensating circuit bells and whistles, taking advantage of driver’s natural impedance and rolloff and selecting the right crossover point.

Another issue is lobbing patterns, which are greatly depending on driver’s sizes, their alignments, crossover frequency and order.
Now, because of room reflections, using the same drivers in the speaker but compounded with different topologies (which are causing different lobbing patterns) can make the same drivers sound very much different in a given room.

planet10

You made me very curious. Can you share what FR and superT you are using and how they are aligned?
I have been trying to use the same setup: 10” Woofer (active 24db/oct low pass @80Hz) + Fullrange (sealed box + active 12dB/oct high pass @80Hz) + supertweeter (6dB/oct @ around 10000Hz)


Argo
 
argo said:
planet10
You made me very curious. Can you share what FR and superT you are using and how they are aligned?
I have been trying to use the same setup: 10” Woofer (active 24db/oct low pass @80Hz) + Fullrange (sealed box + active 12dB/oct high pass @80Hz) + supertweeter (6dB/oct @ around 10000Hz)

Argo,

My current speaker system is quite modest, but very musical, decidely Frugal-phile material. They are a set of BD-Pipes with Radio Shack 40-1197s (not heavily tweaked yet). I have been trying many Ts on the top (physically & freq wise). I am leaning towards some little cones on the top. On each side i have a pr of Peerless 8" aperiodically loaded. More pics and info on my website -- BD-Pipes.

dave
 
argo said:
planet10
You made me very curious. Can you share what FR and superT you are using and how they are aligned?
I have been trying to use the same setup: 10” Woofer (active 24db/oct low pass @80Hz) + Fullrange (sealed box + active 12dB/oct high pass @80Hz) + supertweeter (6dB/oct @ around 10000Hz)

Argo,

My current speaker system is quite modest, but very musical, decidely Frugal-phile material. They are a set of BD-Pipes with Radio Shack 40-1197s (not heavily tweaked yet). I have been trying many Ts on the top (physically & freq wise). I am leaning towards some little cones on the top. On each side i have a pr of Peerless 8" aperiodically loaded. More pics and info on my website -- BD-Pipes.

What FR are you running?

dave
 
argo said:
Currently one Jordan JX92S and one Alcone 10" aluminum woofer.

Does the 10 keep up with the Jordan?

Not decided about supertweet jet - may be some ribbons. But planning to arrange four Jordans and supertweet into DiAppolito with one subwoofer.

In a 2.5 way? You'd be pushing to use the Jordans above 4 or 5k in an MTM.

dave
 
phase_accurate said:
I fully agree with this. Might be worth considering something one could call a 3.5 way.

That would almost be a better description... i have the bad habit of ignoring the woofer in a situation like this. When i think about multi-way speakers i almost always assume an active sub that we can just sort of slide off to the side when thinking about the sat. Even better would be 2.5+sub to make sure that it is understood that the 0.5 section is on the Ms.

So to be clearer i am just taking about the MMTMM. This would be best done as a 2.5, better yet as a bipolar MTM w MM on the backside.

A wilder idea would be to put 1 driver on the face of a square prism. Don't know if this would work out at all (but i still remember being mightily impressed with EPI miniTowers & Towers). I have enuff RS 40-1197s that at some point i will give this a try.

dave
 

argo

Member
2001-02-12 3:35 pm
Estonia
planet10 said:

When 2 sources of the same frequency are futher than a specific fraction of the wavelength of that frequency you start to get comb filtering and you start to lose "focus".

So to be clearer i am just taking about the MMTMM. This would be best done as a 2.5, better yet as a bipolar MTM w MM on the backside.

dave

Fully agree with that. I have tried to take that into consideration.
I can't post the actual simulation graph right now but according to simulation, accounting fullrange driver's natural rolloff and MMTMM alignment and associated comb filtering, causes the on axis response to follow LR 4th order, Fc @ 8000Hz filter response, followed by small bump at 17kHz. Off axis of course is much uglier but it is in the same league with MTM alignment. Oh yeah and design point is 2 meter distance on the sim instead of 1 meter.
Whole purposes behind using this kind of setup (kind of semi line array I would call it - Look at D’Appolito’s Line Arrays at Zalytron) is to use it in large room without loosing loudness and clarity. I have experienced, that Jordans can’t be pushed too hard without distorting. That’s why I tried to increase cone area and decrease driver excursion.
To tell the truth, I am very satisfied with one Jordan per speaker, even without the woofer and super tweet, playing in my relatively small living room but as I have several of these drivers at hand right now, I thought to toy around little with them.
About using fullrangers with their special qualities in 2.5 way configuration is waste of good driver (especially considering Jordan’s price) IMO.

Argo
 
The simulation I talked about in my previous post was made with using modeled, ideal fullrange freq. response graph.
Loaded the actual JX92S response graph into sim setup and got much worse total freg. response: Fc @ 5000Hz, 4th order and following an ugly bump at 13.5kHz just -5dB down above overall level.
So:
MMTMM - project abandoned;
MTM - still possible;
MTM 2,5 -see my thoughts in previous post

Argo
 
argo said:
About using fullrangers with their special qualities in 2.5 way configuration is waste of good driver (especially considering Jordan’s price) IMO.

I don't know. I have been thinking about a Jordan bi-pole (back driver could be run full-out or rolled off for 2.5). No bafflestep and the extra dynamics are pretty compelling. Look at it from the point of view of the amp... you can get away without that extra 6 dB (ie amp of 4x the power). Often worth the couple hundred bucks & bigger box another pr of Jordans would add to the price -- and you gain the benefits of push-push.

And i have had thots of an MTM bipole using some of my collection of FE103A or RS 40-1197s (i have a few of those too). I don't know how it would work, but these guys are somewhat limited in how loud they will go (3W rating) so to get dynamics you need more of them. I figure i can get them close enuff together to XO at about 5k.

dave