Looking for a few....

[dsrviola]

....road signs to get me where I want to go:

In my next project I'd like to make sure I get my vertical dispersion (especially above the speaker) as even as possible. Not necessarily perfect, but better than some of the commercial designs I've lived with. (Matrix 804 comes to mind as one that was particularly week in this dept.) I'm not one of those people that only listens to the stereo from the sweet spot. Actually, just the opposite: I'm rarely in the sweet spot. While I realize that horizontal dispersion is a higher priority, Im not interested in living with a'nother speaker that sounds so drastically different when standing up.

My experience has been that closely spaced drivers, whether mtm or tm helps. It's also been my observation that having the mid (if it's a single mid design) helps with vertical lobing issues above the speaker. I'm not sure if that's because of the type of crossover used with that example, or just a lucky coincidence.

In the case of an mtm, it's my understanding that running the crossover as low as possible also helps with this. I've also "heard" that steep crossovers can help with this as well. Correct me if I'm wrong, but mtm's actually have more lobing issues than an tm, but the issues are consistent from top to bottom, unlike a lot of traditional tm designs that have severe suckouts at the crossover above the tweeter axis.

Something I'm less clear on is the subject of crossovers in phase quadrature. I hear from some that they have a better power response that other crossovers, but I've also read that they exhibit greater peaks or nulls in the off axis than other crossovers: http://www.stereophile.com/content/s...inkwitz-page-4

I'm currently listening to an mtm with a crossover that's in phase quadrature. I'm about to put together 2 different crossovers that exhibit similar (nearly identical) on axis response but one will be in phase quad and the other will be in phase (at crossover). Curious to see how they sound different. I'm relatively pleased with how good the current version sounds off axis - both vertically and horizontally. C to C spacing of mids is about 9.5" and the crossover is (on the high side, I realize) at about 3.2kHz.

I know this goes against the grain of some of the powers that be, but I'm generally not a fan of complex crossovers. I'd rather work with well behaved drivers and keep the circuit simple. No, I'm not talking a cap and coil 2 way (thank goodness I've graduated beyond THAT particular phase), but the simpler the better, especially considering how expensive the xover parts I like to use are. :cry: The tweeter I'm committed to using won't crossover all that low (compared to some), so I'm stuck with going no lower than maybe 2.5kHz.

I'm torn between going with an mtw layout and an wmtmw. Some people say (and this has been - to a degree - my observation as well) that mtm's exhibit a vertically limited soundstage: like you're listening to the sound through a wide slit. Furthermore, that (all things being equal and assuming that distortion due to excursion is a non issue) an tm just sounds more "pure" than it's mtm counterpart.

Anyway...just trying to get some ideas and information to figure out which fork in the road to take. Thanks for taking the time to read through all this meandering.

Cheers!

[Generic George]

Well if vertical dispersion is your primary consideration. I think you want to look at some sort of Line Array. You can't beat a floor to ceiling column of drivers for vertical dispersion. The number you'd be using means that cheap drivers can actually work very well and that crossover points become more flexible.

The major downside is obviously the size. Though cost and complexity can be issues as well.

Parts Express have been teasing us with a curved dsp based line array, using something called Constant Beamwidth Transducer technology. I've seen price estimates in the 2k range, but no firm prices (or really all that much information). Cabinet parts are included for reasons you'll see in the link.

http://www.parts-express.com/cbt36/

[cjd]

Omnipole may also be an option to consider.

As with so much in life, you'll have to balance compromises. You can't avoid them.

MTM will introduce nulls at some point off-axis - I have some measurements of my MTM off-axis that shows this clearly. Those go to between 15 and 20 degrees with the original 2100Hz crossover. The flip side of the nulls is that you get more even dispersion (lobing really) characteristics, and in particular you'll get equal off-axis either way. This may not be a concern for you (or is it? Do you sprawl on the floor to listen at times?)

TM works great but now you have a little more worry about excursion induced distortion on the low end, maybe. Polar response on the vertical access has been documented out there, worth looking at.

If your center distances don't properly account for frequency at crossover you also risk nulls off axis just from this. Shallower, simpler slopes need more wiggle room here, so "simple" crossovers make for much steeper requirements on the drivers. For a 2nd order slope I'd keep center-to-center at no more than about 4.5" - 4th order, 5"... but that's me. And I've broken that rule.

I think I would recommend at least a WTMW layout - I've found that you lose some of that fullness if the woofer(s) end up down low. Even the big WWMTM's I have fill out in the lower regions when I get into the vertical sweet spot, and on those, it's tough to avoid - they're big enough, with a pair of 7" mids (10" woofers) and a ~1600Hz crossover to the tweeter, that unless I'm right up against the speaker I'm well within the window for good response, vertically. Steep slopes too.

Which brings me around to that particular topic: When you spend on the crossover, you can pick drivers with a narrower range of perfect. Or you spend on the drivers so you can save a buck on the crossover. However - what type of music do you listen to? I'm guessing you've avoided the low-bitrate-MP3 problem, but maybe that's what you use. Rock? Classical? Rock in particular inherently builds in so much distortion as part of the desired sound profile (guitar amps intentionally push tubes into distortion for effects - this is the type of thing I'm referring to) that you just don't notice if the crossover doesn't adequately manage rough response problems in the drivers. Classical? Well, for me, as a trained classical musician, I can hear those tiny little things that aren't right. And while they don't get in the way of a quick listen, and certainly not if I'm just wandering about the house with music on... if I'm sitting to listen, they start to creep up on me.

Neither way is wrong, just a different choice of compromise. Though in this case, as I mentioned before, you may get more off-axis null on an even order crossover with lower slopes. That may be an argument for a properly done 3rd order slope.

FWIW I'm currently embarked on a project that I hope *may* let me experiment with 1st over electric circuits. So I'm not married to 48dB slopes.

[JonMarsh]

Some quick thoughts-

Don't have time for a long response at the moment, but some thoughts that may be helpful-

Phase quadrature can help an MTM with vertical dispersion substantially. It's inherent in a Butterworth 3rd order. Google online, you should be able to find some comparisons of vertical polar dispersion, or model it yourself in LspCAD or SoundEasy.

Phase quadrature is what Curt uses in the well regarded Statements, and it makes a big difference, I believe, considering the driver spacing and crossover frequencies he's chosen.

If you're investigating different crossover functions in general, pondering your options, one I think you may benefit from looking at closely is the Duelund concept for three ways.

Last, if you think you want an MTM driver configuration, but are concerned about the vertical dispersion issues, then consider going with a 2.5 way approach- this give you the best of both. I'm doing it in the midrange of my current build.

Image not available

I'm in agreement with much of CJ's comments, except the part about rock and being more tolerant of system defects- for me, the dense midrange sound seems certain to excite issues- in fact, I have a number of albums I like musically, but are also stress tests for DAC's as well as speakers, and finding a setup that keeps everything clean in the midrange through presence region is no easy task. I certainly agree that those issues can stand out for music that's acoustic based and for which you're familiar with the original sound.

Last, using a lot of crossover parts to beat a design into submission isn't often a good idea, if only because to avoid those parts making their own sonic contributions, they wind up being pretty expensive themselves! Somewhere there's the right trade off- another reason I've been a bit enamored of waveguides and careful implementations, as it's enabled reducing the driver power handling for a given SPL, and reducing the crossover component count for a given acoustic target.

Good luck with what you seek- and keep us posted on the journey.

[cjd]

To be sure, rock will excite the same system issues, but people that don't live much outside of rock may not pick up on them. As I say this I have my brother-in-law in mind: he seems to listen exclusively to thrash metal... all noise to me, I'm not very knowledgable but he'd probably tell me I have it all wrong.

Such music better at masking distortion issues to a less trained ear. That's a way of saying, you, the listener, may not care as much that you crossed over high enough that any distortion peaks aren't all that attenuated... that's all. It's the listener that's more tolerant, and then only maybe.

So crossing a 7" driver at 2500Hz may work just fine. I've seen and heard, for example, the Seas W18EX crossed 2500 4th order despite a notable F3 distortion peak at ~1700Hz and, of course, the crazy 5k peak that shows up everywhere.

I've found it is easier to make people that just want to rock out happy, than it is to make someone that prefers chamber music happy. Both of which I find are steps along the way to getting a system to handle orchestral works properly.