DIY HT Design Coax or Seperate Mid/High?

In general principle I like coaxes and have owned others in the past, mainly a lot of Tannoys. So I would vote yes to a coax based design, but it would depend upon the alternative's performance and the listening distance.

I think it would be an awesome design with one, but like you noted the biggest hurdle is the budget so I voted for the standard config

As I've stated in other threads I'm hoping someone would develop a poor mans catalyst without amplification and a passive crossover. If you go with a coax you are one step closer to something I could get interested in. I'm also closely monitoring the CD/W designs under development. If I wanted a speaker made with the same old mix of T/M/W I would build the Statements, Khans, or any one of the other engineered and tested designs that have been throughly documented.

I can't explain why I am only interested in something that is different, perhaps it is the misguided belief it might be somehow better mixed with a spirit of DIY adventure.

You are not alone, I have also posted this opinion during this process.

I'd like a high efficiency coax-based design to use for surrounds to go with WG fronts.

I'm partial towards the B&C's, as I've read nothing but good things about them.

The wrinkle is the XO. B&C sells one for $60 I think, but I don't know if any of the good things I've read are with it.

"Another noteworthy Coax, that MIGHT fit the budget. $175.00 is workable, i we can find some independant performance data.

http://usspeaker.com/homepage.htm"

That's just the home page; which driver do you mean?

One advantage of coaxes that hasn't been mentioned is symettrical polar response. In practice, that means one can stick her/his speakers above an LCD without problem, or without changing the design compared to one designed for use at ear level.

Unless the mains use the same, timbre matching gets flushed.

I've never met a co-ax I liked.

Here's Beta 10CX with a bunch of different drivers:

http://www.diyaudio.com/forums/showt...=&pagenumber=4

The answer is "No."

"I've never met a co-ax I liked."

Mark Seaton's acclaimed Catalysts uses the B&C (active XO though), and it's also used in a highly regarded pro speaker with hifi tendencies that I can't name.

There are "issues," even with the new one:

http://www.bcspeakers.com/PDF/PRD/8CXN51.pdf

Why'd they grab 16 kHz for the polar plot?

The BMS ones intrique me because they've got a "proper" horn.

Early pics of GPA's 12" Altec Duplex reissue/upgrade also looked promising, but I don't know if he stuck with that design....

I sent Auger an 8cxt to test so I guess we'll find out.

Mayhem, I sense an edge in your comments that leads me to believe you are reaching your tolerance level for this noble but frustrating path you have chosen to build an AVS unique speaker design. From the beginning it was all about a survey based consensus building model of product design.

Unfortunately the reality is that some members of your voting audience are hacks like myself who rely on the heavy lifting of others to do the really technical engineering.

Pay attention to the true experts, keep soliciting input but build what serves your needs. If you try to please us all it will drive you bonkers. If it turns out great you will have some followers.

"Those craving ths Coax really need to explain the benefits to the group"

For the pro-based drivers:

1) High efficiency

2) Low power compression

3) Clean at high SPL

4) For the B&C's at least, largely CD (constant directivity) behavior which gives even power response and may bring the same benefit of wide sweet spots to surrounds as given by cross-firing CD L/R's.

"I sent Auger an 8cxt to test so I guess we'll find out."

"a 6.5" high FS cone with a CD crossed 3-3.5khz makes the most sense."

What's FS?

Also, I think 3-3.5 kHz is too high to properly match directivity at XO; I think the B&C XO is at 2.2 kHz.

Great!

Resonant frequency (of the LF driver.)

The best reason is for the physical alignment of the tweeter-midwoofer. You get around all the problems of having the tweet-midwoofer in two different locations on the vertical plane. You get less lobing on the vertical axis, and better power response.

The biggest problem is as Noah mentions. Most of the devices on the market have some design issues. The tweeter/midwoofer directivity doesn't match up at their crossover points. The response of the tweeter, midwoofer or both tends to suffer from the environment it is forced into by the physical constraints. You are forced into some compromises on transducer design based upon the physical limitations.

Overall though.... I think most of the negatives can be overcome. I just need a pile of money to do so. ;-)

I gotta agree with Mayhem. If anyone were to look at that woofer response and suggest to you that you should spend 100+ for it and expect to cross it over above 2khz, most designers would say no way. The tweeter response is much worse at another $100. I doubt anyone whould buy those two drivers were they sparate, with that kind of response at that price level, crossed where they would have to be crossed. These are horrible looking on axis responses. I believe first arrival sound should be first on the priority list and that means excellent on axis response should be a top priority. Why is anyone willing to sacrifice that first arrival sound quality in favor of a bit higher sensitivity and the ability to play to an unrealistic, ear damaging, peak spl levels? I wonder why you don't see Geddes using coaxes in his Pro driver/constant directivity solution? As Mayhem said, it can't be a good thing to have a constantly moving waveguide or a waveguide whose design is secondary because it has to act as a driver first. To me, there is so much compromize in the design of the coax, as can be seen by observing BOTH drivers responses, that it swamps any possible benefits it may have and compromizes the design of both drivers. So you end up with two very substandard driver responses just so that you can make them into a coax? I don't get the attraction, personally.

There is a reason why MT or TM is predominant, because it works. You can use a pro midrange, which is pretty flat out to 4khz. The SB29 Ring radiator tweeter looks to be have a sensitivity in the 93-94db range and very good low end performance, flat down to 1khz, for a nice diver overlap. Mated with a pro woofer you should be able to get in the 92-93db range, depending on applice BSC, have very low distortion and be clean at all but insane levels, IMO.

If you want to increase the low end output of the dome tweeter, potentially increasting its power handling, increase directivity and create some pattern control above 3khz or so, you can mount the tweeter (less flange) from the rear of the baffle in a shallow waveguide. A DIY waveguide, at 3/4" deep with a 3/4" roundover you can get a boost in the range of 4-5db at the centered around 3-4khz and get some directivity benefits etc. Even Geddes admitted that if he were to use a dome tweeter, he would mount it in a waveguide.

Below is a study of the Peerless HDS in various depths of shallow DIY waveguides, but first is a pic of the final WG mount (in Lexan and wood, with open baffle mid, not suggesting that for this project). The waveguide is created by nothing more than a 1-5/16" forstner bit and a 3/4" roundover bit, so it easily doable by most DIYers. With the tweeter flange removed, you can decrease the CTC distance of the mid to tweeter which improves vertcial polar response. Put the waveguided tweeter on a sloped baffle and you can align or come close to aligning the tweeter and mid AC in the vertical plane as well.


HDS mounted Flush: 0-15-30-45-60º



1/2" deep WG, 1/2" roundover: 0-15-30-45-60º


1/2" deep WG, 3/4" roundover: 0-15-30-45-60º


3/4" deep WG, 3/4" roundover: 0-15-30-45-60º


3/4" deep WG, 3/4" roundover, with 1/2" deep 45º chamfer after, sharp edges sanded round by hand: 0-15-30-45-60º


3/4" deep WG, 3/4" deep 45º chamfer, sharp edges sanded round by hand: 0-15-30-45-60º

The above is the primary reason for using a coax, along with the flexibilities these qualities afford. If you're used to looking at dome tweeters or a smoother responding horn over a limited angle, the high frequency will look like garbage.

On second thought, Zilch & dl are right, the response is worthless, nothing to see here, please move along... Those things can't possibly sound good.

I'm not sure how to interpret this. It comes across like you are taking your ball and going home.

First, anyone who doesn't use not just the exact same drivers, but the exact same drivers in the exact same cabinet at the exact same elevation for all three front speakers just doesn't care about high-fidelity reproduction. So who cares what they use?

A big advantage of coaxes, as I mentioned above, is that because of their inherently symmetrical polar response, one can mount them higher than one can MT's, to clear LCD's or other things, and not suffer sonically for it. (Yes, one can also design an MT to work at any given height, but then you move and need a new placement, and have to go back to the drawing board. With a coax one just modifies the stand height.)

You're basing a dislike of coaxes generally on the basis of a cheap Eminence? Really?

I hate JBL. I once heard this laptop that had JBL speakers, and it sounded really tinny.


That's nonsense. Tannoy, for instance, has long used fairly low slopes, I believe 1st order/3rd order asymmetric on my the DMT series.


2.2kHz is probably still too high. I think one should be thinking of 1.4-1.6 kHz with an 8" driver. That's what Tannoy does, and their Dual Concentrics are still, of what I've heard, the best of breed.

You just smooth it a little more until it looks good Mark. ;-)

I don't really care about top octave jaggies. Once they start moving down below 10K though I become progressively more concerned.

I'm convinced it could be done without any tradeoffs in terms of smoothness of response. It would require tooling a lot of parts though.

The idea is really attractive to me for home audio sized devices. It certainly makes designing different configuration speakers easier as you can turn them anyway you like.

It's true the FR looks awful, but the integrated response w/XO is much better, and Earl Geddes has noted that B&C's data is accurate vs. the sugar coating you get from a lot of other mfgr's.

Also, I'm counting on auto EQ to smooth out the significant bumps, and a 100 Hz XO to the mains should keep cone excursion under control.

Anyway, I bet if anyone measure's the raw FR of their surrounds from any one listening position that most of them are pretty bad because of the reasons Kevin noted, unless they're far enough away to minimize angle differences between drivers.

That isn't that bad either. Sure... it isn't perfectly smooth like some small domes but it is usable. You can easily get that +/- 2dB without any trouble. The power response is more important than an extra +/- 1dB of linearity in a non-smoothed curve.

Presumably, you have some measurements to post in the interest of furthering this discussion.


They were mentioned as an option here. We have found them suitable for PA at Burning Man; Eminence themselves have indicated "appropriate" application.


Best Buy Insignias are coaxes....