[BigRick]

Does anyone know any money and by money I mean spectacular speaker cabinet builders in St. Louis.  Or is anyone willing to stake their reputation on a build enough to take cash upon completion of a build?  Otherwise would anyone be willing to lend me some space in there wood shop, in I don't know 8 months to a year?  I am in the process of designing an active crossover full-range tower speaker with a full closed cabinet with 4 ports and 2 12"peerless woofers and 2 5.5"peerless mid ranges and a fountek ribbon tweeter,  If anyone has any advice on where to get this cab built or suggestions on driver selection and where to buy that would be great.  Any help would be greatly appreciated.

[G]

Does anyone know any money and by money I mean spectacular speaker cabinet builders in St. Louis.  Or is anyone willing to stake their reputation on a build enough to take cash upon completion of a build?  Otherwise would anyone be willing to lend me some space in there wood shop, in I don't know 8 months to a year?  I am in the process of designing an active crossover full-range tower speaker with a full closed cabinet with 4 ports and 2 12"peerless woofers and 2 5.5"peerless mid ranges and a fountek ribbon tweeter,  If anyone has any advice on where to get this cab built or suggestions on driver selection and where to buy that would be great.  Any help would be greatly appreciated.

Sounds like you are building a old Infinity SM150. If you find someone please post his name and number!

[dwarfed centipede]

Are you pairing 2 woofers with 2 mids?  I would suggest 1 woofer and 2 mids for balance purposes.  Or do 4 mids and 2 woofers.  Just my 2 cents. 

I wouldn't call myself an elite speaker builder, but I enjoy doing it and would be willing to build this for you for minimal cost.  If you want, you can come over and see some of the speakers I have built.

[BigRick]

I would like to come by and see that when I move back to good ole STL.  I would also like some more feedback on the 1 woofer two mids.  Through extensive research I have seen overwhelming support for the wmtmw design for loudspeakers.  I will be using an active crossover so leveling shouldn't be an issue between the sections.  The tweeter I am looking at is 96db efficient so it will have to be leveled down of course.  But I want tons of dynamic range on the upper and lower frequencies, for creschendos and the like, most of what I listen to is fine art music, whether it be classical, romantic, experimental alternative rock. 

So far box tuning has the Woofer tuned to at -2 of around 20hz with a possible peak SPL of 111 Db with 200 watts on tap with the 2 woofers and that was one of the design goals.  I am however considering switching to a pair of SEAS for the critical midrange component, possibly something more efficient than the peerless but I wasn't sure if voicing would be a concern there. 

Thanks for any more input, I am in the modeling and late design phase at this point, then I will work on fun box design and modeling, then the tweaking of the final components and sourcing, then making the design for a not only functional cabinet but a pretty cabinet, then of course buying building and finishing.  I want a 50k performance for 3-4k price.  A large chunk of that 3-4k is going towards the active crossover for a chance at some trial and error.

[dwarfed centipede]

I think the active crossover is a great way to go.  After listening to Bob's Behringer, I really want to get one so I can tweak my crossover points.  I guess the 2 subs will work, but I was worried about it overwhelming the 2 mids.  I like seas speakers, I think they are pretty efficient too.  If you can get a few woofers over 92dB effic, Maybe the wmtmw design will work alright.  Two 93db woofers would give you 96db.  I just havn't seen many designs in that arrangement using 12" woofers.  Are you going to compliment these speakers with a separate subwoofer as well?  Or do you want all your bass coming from this cabinet?  If you want sub 20hz, these speakers are going to be rather large and heavy.  But I think you already knew that.

[G]

The speaker you are talking about is going to be huge. I owned a pair of these back in the day. They had a sensitivity upwards of 98 dB 1w/1m. They used a single 15" for the lows and they definitely didn't need any help from a sub. 75 watts will drive these into the ground so you can focus more on quality than quantity.

[BigRick]

Yeah, size is not a compromise I have to worry about my plans so far are for 11ft^3 each for the woofers and the final speaker should be 60-70" tall with the tweeter in the middle right at 34" ie ear level for my seating position.  I will build with 3/4" mdf throughout and feed each speaker up to 600 watts through the active crossover.  Price and width are my only concerns as long as the speakers aren't 10ft tall and 1 ton each.  This will be a long build process as I just bought some really good speakers but I'm at the point where I am not willing to pay for a factory built "no compromises" speaker.  I still have some time to make tweaks I probably won't buy any parts for at least a year but guys on this forum are always such valuable rescources.

[BigRick]

after a little more modeling 4 mid ranges would not be a bad idea, and then I could use the back area behind the mere 1 ft^3 of air need for all four mids to deliver their flat middy goodness as extra air space for the woofers making for a less meaty speaker with a more dynamic and effortless midrange.  The 4 mid-ranges gave me headroom of 113db while 2 only gave me 109 while the woofs have 111db.  Seems legit in concept, with this tall of a speaker would a curved front baffle be necessary.  I am not really sure how to model phase response and edge diffraction, the active crossover has a 2ms phase adjustment but I prefer to leave that off if possible and have the speakers as naturally time aligned as possible.

[dwarfed centipede]

Iv noticed on some good engineered speakers that they recess the mids from the woofer, and the tweeter even farther back.  I think the goal is to get all the voice coils on the same verticle plane.  I've not researched this that much, if you look into it please link the articles you find.

[KenSeger]

I've noticed on some good engineered speakers that they recess the mids from the woofer, and the tweeter even farther back.  I think the goal is to get all the voice coils on the same verticle plane.  I've not researched this that much, if you look into it please link the articles you find.

I've fiddled with time alignment with the Bozak tweeter and midrange.  I measured the distance from the voice coil ends to the grill cloth, did some math ala a triangle with two equal sides and adjusted the distance between the speakers, my distance to the speakers and the angle of the speakers to each other.  It did make a noticable improvement.  That would be the "moving the mountain to Muhammad" method - Concert Grands are NOT lightweight.

A more straight forward method would be this:

-----------------------
Ken,
 
A few comments are appropriate.
 
Each driver has an "acoustic center" which is the plane from which the sound appears to originate.  Typically, it's located near the end of the voice coil former, but there are exceptions.  In any case, it can be measured.
 
So a "time aligned" speaker would have all the acoustic centers equidistant from the listener.   That is, assuming the crossover doesn't introduce some phase shift which moves the acoustic center.  Typical driver arrangements for "time aligned" speakers have been stepped, sloped, or curved baffles.  Vandersteen makes "time aligned" speakers and in the 1B uses an angled 8" and a vertically mounted tweeter in a kind of "'free space" array.   The Dahlquist DQ-10 was time aligned and mounted the individual drivers to small baffles which were then affixed to a screen which curved up and back.   The drivers with the deepest acoustic centers being mounted low and forward, and the shallowest mounted back and higher.
 
Something to think about, however, is that the drivers are only time aligned if you are on axis.  As you move off center the alignment is lost.  To see this more clearly, think of being 90 degrees to the cabinet.
 
Phase alignment is something different.  If two drivers have their acoustic centers 1 wavelength apart (in depth) at the crossover frequency and they are wired in phase, then the output from the two will be in phase as the second wave from the closest acoustic center will be in phase with the first wave from the farthest.  That's not time aligned and it will show up on transients.
 
There has been a lot of discussion about the out of phase mids in Bozaks.  That seems to have been an attempt to get phase alignment between the woofer and midrange for better transient behavior.  The idea being that the second half wave from the mid will be in phase with the first half wave from the woofer if they are wired out of phase.  The Yamaha NS-1000M used this trick.  It won't work with a 400 Hz crossover, but the distances between the acoustic centers are closer at 800 Hz.  Since the wavelength of sound is about 1.4 ft at 800 Hz, a half wave is 0.7 feet or 8.4 inches.  Not equal to the distance between their acoustic centers, I'm certain.  Further, we are talking about the distance to the listener's ear, and since the mid is higher mounted than the woofer, the path length will be longer for the woofer sound, adding to the offset.  But that added offset will be small if the listener is seated well back.  At least that's the theory.
 
As I wrote that, I wondered if some whiz-kid didn't get the woofer phase aligned with the mid from a microphone on the midrange axis, at some close distance where the path length difference would be maximized.
 
So what you are looking for is time and phase alignment at the listening position.  Usually this is implied by the term "time aligned" but it is not a necessary condition, so you do need to be aware of the difference.
 
Earlier, I mentioned crossovers.  The only crossover which is phase and time aligned is the first order.  Higher order crossovers add delays that have to be compensated for. For that, you need to be a Siegfried Linkwitz. (inventor of the Linkwitz-Riley crossover filters).  Since Bozaks use first order filters, they are a good candidate for time aligned systems.  All you need to do is find the acoustic centers and design a cabinet that has the right mounting distances for the driver, so that the path lengths to your ears is the same.
 
One of the problems with first order crossovers is that in the overlap region, where both drivers are contributing, they form a phased array that has directional properties different from the individual drivers.  Depending on the offset between their centers (here we are talking about the center distance between the drivers, not the acoustical centers), the directional lobe will be narrower or wider and will be steered off axis.  For drivers mounted in the same physical plane (a flat baffle) a first order crossover steers the lobe downward.  Mounting the tweeter back so the acoustic centers are aligned gets the lobe up on axis.
 
I have run some experiments on time alignment.  There is a noticeable increase in clarity when the various parts arrive together.  I made a set up where the tweeter could be move for and aft above a fixed woofer.  Subjects were seated on the tweeter axis.  The subjects were blindfolded. (I used my teen age daughters, both of whom sang A Cappella) They were told to say something when they heard a difference.  The music was started and I began moving the tweeter along a strip of masking tape.  When they said "stop", I marked the location and asked them why they said stop.  The answer "it was clearer".  All of the marks where they said stop were in a 3/8" space!  They were repeatable to within +/-0.19"!!!   Then I sat them on the woofer axis and repeated the test.  Again, all the marks were within a 3/8" space, but at a different location a little further forward on the tape.  When I measured the path lengths from the end of the coil former in the two locations, they were time aligned in both.  Crossover was at 3 kHz where the ear is most sensitive.  The wavelength of sound is about 4.4" at 3 kHz  so they were getting the alignment right to +/- 15 degrees or so.
 
This made a believer of me.
 
It does appear that sensitivity to time alignment is a maximum in the mid frequencies and is less so at low frequencies, perhaps because the wavelengths are so long and the ear is much less sensitive.  So, I'd worry first about the mid/tweeter alignment, and then use as low a crossover as I could.  Front mounting the mid would help a lot.
 
The practical realities of getting several inches of offset between the mid and woofer is tough to do without a step that can cause reflections of midrange energy and consequent bumps and dips in the response curve.  You might be better off not aligning the woofers with the mid.  A way out of this would be a biamped system using a crossover that has variable delay (like a Behringer digital) so the position of the mids can be moved back electronically.
 
I hope that clears up this subject somewhat.
 
Jerry
--------------------------------
Ken

[BigRick]

I am thinking of buying a Behringer 2496 active crossover and it has the timing delay.  Does the time delay introduce any noise to the signal and thus distortion or is the delay strictly a mechanical process?  As far as the phase goes am I getting this straight, I just need to make sure the acoustic centers of the drivers are 1/2 the length of the crossover frequency wave? 

Basically, the underlying question is, can I get essentially the same effect from a curved baffle as an active crossover with delay taken into account, or would the crossover introduce noise?

[Bob in St. Louis]

But how do you know (or find out) the acoustic center of a driver? 

Bob

[BigRick]

Yeah, that is what I was wondering too.

[KenSeger]

BigRick,

Two different subjects as far as I can see.  The arguement for phase inverted mid-ranges in a first order (6db/oct) system is addressed by the half-wave placement (side to side or top to bottom) whoopla.  Time alignment in my mind (placement front to back) is an entirely seperate subject.  At least in the Bozaks (which is what Jerry Crum was addressing) the phase inverted midrange yields such a mucked up sound stage that I really don't care what the frequency response is at the crossover point. Dips or peaks I can stand, getting seasick from minor head movements, no thanks.  Half of the art of hi-fi is finding out what errors in sound (or omisions) you can live with and what errors or omisions you can not tolerate.  Moving your head several inches from side to side even over 15+ feet away from the speaker changes the sound stage so radically, the ideal listening position is measured in inchs instead of feet!  If one had the equipment to measure these lobes I think the graph would look more like a spread pattern of a sawed off shotgun than anything associated with hi-fi.  Since I personally have no use for phase inverted drivers, it's on to time alignment.

Bob,

The only way to measure what Jerry Crum refers to as the acoustic center (front to back of the driver) would be a very sophisticated measuring system with single fullwave pulse going into a driver with the microphone at a set distance to that driver, with the original (electrical) pulse being time delayed so that the acoustic pulse would line up with the electrical pulse on a dual trace osciloscope.  That would then be your reference point, or more properly, reference distance.  You would then take the next driver (above or below in the crossover, so starting with the midrange as the reference distance makes sense) and move it back and forth until you got its pulse to match and that would give you its distance.  Repeat with the third driver (for a three way system). *OR* you do the whole thing by ear.

Start with the woofer and add a midrange that goes all the way up from its crossover point, moving the midrange back and forth until "it sounds right".  Then, with the midrange with its correct crossover, do the same thing with the tweeter and move it back and forth until it sounds right.

BigRick,

Meanwhile back at the crossover - to my prejudical way of thinking, anything extraneous that you can remove from the sound path is "a good thing".  Therefore I'ld be all for prefering time alignment to be done by physical placement rather than on the backend with a digital time delay.  Of course if somebody gifted me with Klipschorns, two seconds later I'ld start pricing two time delays and tri-amping.  As for distortion products from the time dalay, I've never heard a Behringer, but I've been told by ears I trust that they are very good.  I will say this about this entire subject, if one considers the old Bang & Olufsen Phaselink or uni-phase speakers as the most complex attempt to neutralize time and phase alignment in higher order crossovers with the phaselink driver operating ONLY at the crossover point (ie. it's frequency response looks like a mountain peak), one can draw solace from the fact that they gave up the concept in 1980.  Analog crossovers are going to have limitations, period.  A digital crossover could be designed that has an infinate slope, ie. woofer flat out to X Hz with the midrange flat from X+1 Hz to Y Hz and the tweeter flat from Y+1 Hz for a nice little tri-amped system and no quibbling.  

Whew, makes you want to chuck it all, buy a pair of Lowthers or Fostexs and be done with it, eh?

Ken

[BigRick]

Yeah lowthers crossed my mind when I started my design but personally I'm not a big fan at rolloff in both directions from the get go, my goal with these speakers are huge headroom, relatively flat frequency response smoothness not so much an issue, as few speaker design errors as is humanly possible, by a first time designer, and a full range sound to my ear 23hz-13.5khz.  Not too much to ask for.  I'm keeping the things that I would not be able to design well with my level of expertise adjustable, such as crossover.  Beyond that, I'm having tons of fun and learning a lot.

I'm thinking at this point, after hearing the experiment with your daughters, to, at least in the beginning, put all of the drivers in their own separate cabinets so they can be tweaked to my liking then building a new box that makes it all work.  Or just keeping them separate for infinite tweakability.