Hey guys...we need a little rallying here... - Page 150

Same here... I want to build a sealed shallow cabinet. I haven't done much modeling but need to find the perfect driver that is not too expensive and works ok with a small enclosure.

You can be more space efficient if you don't make a rectangular cabinet. One example.
Putting the horn axis at ear level, it's not hard to get 60 or 80L (net) with even a rectangular cuboid that's fairly shallow.

Are you thinking of something like the MK S-150THX which the L/R's are squared up to the wall in the back, but have angled baffles, so you don't need to toe them in?

When you say there isn't nearly enough, does that mean one more brace or 3 times more? I'd say I'm pretty near because there's not much room left for more bracing. When I first laid out the box I figured on 2 shelf braces and one vertical, but I changed my mind. The outside is 15" wide. Inside is 13.5". One brace leaves only about a 7" span. That's overkill even for a LMS 18" subwoofer. The box is 22" tall, 20.5 on the inside. So there's less than a 10" span in that direction with one shelf brace. And a 1.5" thick baffle. Seems pretty good for this speaker design.

I can get an extra horizontal brace put in them if people think it's necessary. But I'd say that would make them more braced than any speaker I've ever seen, commercial or DIY. Even the JBL pro theater speaker boxes that use two 15" woofers only have one 2x4 that goes from front to back and side to side. And they only have 1" thick baffles made from particle board with unbraced spans well over 20" for dual woofer boxes.

I certainly understand the desire for overkill. On the speakers that use the $200+ woofers, I can get another horizontal brace added if needed. They're going to be a different box design anyway. But there's no real need for an extra vertical brace to make the internal spans only 4" wide. I've never seen any box made like that.

One thing to keep in mind is that cabinets that are producing midrange need MORE bracing than subwoofers, not less.  With a subwoofer, all the output is way below the wall resonances, so bracing is really hardly needed at all unless the walls are so thin they'll start puffing out.  But for woofers/midranges, the walls want to vibrate right at the same frequencies the driver is shaking them at, so subdividing the walls pushes the resonances higher in frequency, hopefully up to where the waveguide is working and won't be vibrating the box.  I've heard a rule-of-thumb of getting down to 6" expanses between braces, but I don't know what that is based on.  But in general, getting the resonance higher (and tougher to excite) is good if a speaker is doing more than sub frequencies.

I am assuming that your name designers will specify specific boxes for their efforts. You should offer those as flat packs and possibly baffles only. In addition, offer a couple of generic box flat packs and baffles. As far as roundovers, I am again assuming that machining a really large roundover on a double thickness baffle would not be practical or cost effective. In that case, forget roundovers on the baffles.

Erich, I'd have to disagree. I'm not a huge subscriber to the "panel resonances are the end-all be-all" train of thought, but at the same time for a high output, large panel'd speaker that will be operating up to ~1khz, it's sensible to push any resonances up high in frequency and reduce their magnitude significantly. An unbraced 10" span would probably resonate right in the lower midrange. I'd try to cut it down to between 4 and 6 inches...

Personally though I feel large sheet bracing adds too much weight without sonic return. I'd say mass loaded vinyl and closely spaced dowel bracing are the way to go.

Here is the bracing used in the B&W . Crappy speaker, but very non-resonant construction.



It's not thick wood, but it's very closely spaced.

Remember that most boxes are not capable of 110+db in the lower midrange.

This box or that?

Our answers should inform and lead to the "better" aspects.

Working with Martin King's worksheets tells me there are any number of suitable designs. One aspect to reducing internal resonances is locating the mid-woofer as close to one end as possible. Pushing this further, if that end is sloped, say 3" across the depth, the effective distance from the center of the driver to the end of the enclosure is 1 1/2" less.

This may not seem like much, but it reduces the first resonance by half.

It is the reason I choose to locate the waveguide outside the box.

Now, with regard to panel resonances, I consider the mid-point of panels to be the most likely to vibrate. I choose to add mid-panel stiffeners that are diagonal in orientation and cross the mid-point of the largest unsupported panels.
Edited by bg40403 - 7/1/12 at 10:25am

Question forthe experts...I am trying to design the shortest box to fit the seos and a 12" woofer. Would protruding the woofer onto the waveguide by about .5" cause any problems with the waveguide doing its job? In other words I would route a radius into the waveguide to fit the woofer. I measure about 1/2" on the waveguide edge prior to it "Turning" down.

I wouldn't be too concerned.. Troels graveson did something similar in his DTQWT:

If a shelf brace isn't used and I just slide in 'stick' braces, it's worse. If you take two 12" x 12" pieces of wood and put a stick brace across them, 95% of the panel still has 12" spans. A shelf brace spanning all the way across at least cuts the number in half right?

The picture of the B&C has the vertical braces not touching the inner panels. It looks like the arched portion touches the sides, but not the verticals. So it leaves a panel that wraps the entire circumference of the speaker with no brace.......which is probably longer than the ones I had done.

Trying to brace a cabinet every 6" (in all directions) is going to make for an extremely complex, heavy build. Putting in 2 verticals and 2 horizontals will divide it up smaller than the 6" mark.........but from front to back there will still be panels that are 5"x16" (or whatever the cabinet depth is). To divide the depth of the cabinet would require 1-2 more vertical braces going side to side. Then we'd have some type of crazy matrix Rubic's Cube that would be nearly impossible to glue up without a significant amount of sanding and cutting. If just one panel is glued slightly off, it wouldn't fit perfectly.

I certainly don't mind putting an extra horizontal brace inside the cabinet. But I'm not sure what can be done from front to back to divide the span of the cabinet's depth in half. The B&C speaker posted above does not have that either. I don't think I've seen a speaker with that type of bracing to divide every panel into a true 5" or 6" square.




A CNC machine is extremely accurate. But the weakest link to putting it together is the end user. If that type of bracing is glued just a little bit off square, nothing else will fit from that point on. No problem for many DIY guys......they just go cut another piece of wood, or cut each brace as they go to make sure it fits. But for people buying kits....there is no other piece of wood to cut.

What are the thoughts on the attached speaker design for a center channel? What does this do to the directivity control and is it as important in a center channel for HT duty?

You've got a big horizontal discontinuity from the increased CTC spacing (you may as well have used a round waveguide at that point) in the crossover region. This will be right in the midrange and since you have two horizontal ears, both your ears will get different phase summation unless you find some magical sweet spot with your head in a vice. Centers IMO need to be coherent and dissappear more than any other speaker, because they're mono and they're directly in front of you. Get the center wrong, and you'll find that it sounds a lot worse than a phantom center.

I think something like Bill's "Malcolm" design just makes way more sense than what you showed, and even then i'm not sure it's as optimal as a true point source (like a coaxial or synergy horn). I really hate mono listening because it's really revealing of speaker flaws - and a center speaker is fundamentally a mono speaker.

When it comes to directivity of a center speaker, remember that it will be physically 7+ feet away from lateral boundaries, so narrow horizontal dispersion is not as necessary as much as just getting an even power response. Arguably wider dispersion is better for a wider sweet spot for a forward firing mono speaker. The most offensive reflections are vertical, and again the speaker you posted has wide vertical dispersion (the seos-12 for example would does not have any vertical pattern control below ~2khz IIRC; any vertical narrowing in the final speaker response is a product of the crossover, and in this speaker it would just be awkward)

The solution for a center isn't always identical to the mains.
Edited by Eternal Velocity - 7/1/12 at 12:24pm

re'I certainly don't mind putting an extra horizontal brace inside the cabinet. But I'm not sure what can be done from front to back to divide the span of the cabinet's depth in half. The B&C speaker posted above does not have that either. I don't think I've seen a speaker with that type of bracing to divide every panel into a true 5" or 6" square

What you can do is put a dado down the center of each side panel (or at the 1/3 and 2/3 points of wider panels) 1/4" deep and glue a 3/4" x 1.5" to 2" strip in them. Then notch the shelf braces to fit around them. When I do this I make the notch about 1/8" overside for 1/16' gap on each side, then I fill the gap with PL premium when gluing the box up.

Jack

The shelf braces couldn't be slid into the box after the top, bottom, and sides were glued together.

There's no doubt that we could do some crazy bracing, but the boxes have to be very easy to assemble. Most people that build boxes all the time probably won't be buying these flat packs. The flat packs have to be designed to fit together with no major hurdles. Things have to slide together and be squared up easily by people without many tools.

One way to *easily* do the vertical braces midway through the depth of the box would be to cut stick braces that could be glued between each shelf brace after it was put together.
Edited by Erich H - 7/1/12 at 4:07pm

Have you checked out BWaslo's Malcolm center design: http://www.diysoundgroup.com/forum/index.php?topic=11.0?

Yes I have and I really like that form factor. Frankly I would love to build an LCR in this form, but I have 3 AE TD12M's coming and want all three identical.

The picture I provided was the Elemental Designs Cinema 12 center. All reports that I have heard are very good, so was just curious what the thoughts were on the "Theory". I have a 20" height limit for the center.

You could probably get the box down to a 20" height if it was really needed.

A couple thoughts/questions......

When positioning waveguide on the baffle is there any rule on distance from top and side edges? If not is there a recommend minimum distance?

Any reason to use rope caulk or other dampening on the SEOS?

FWIW The 6" rule for bracing ive heard refers to an unbraced panel of .75" MDF. Thicker material or different substrate would resonate slightly different.

Dowel bracing might be an option for dynamic box sizing and the end user could cut to length.

Are there no measurable artifacts from not recessing the waveguide? No diffraction off woofer frame either?

Unless you're following a specific design then the waveguide position won't matter much. Depends on the design.

These seem pretty robust. I doubt any dampening is required, but can't hurt.

I'm not sure recessing the SEOS changes the response via diffraction but with woofers, they're usually so directional near the wavelength of the woofer frame that recessing doesn't matter.

The waveguide is pretty shallow and thick and you really don't need to worry about rope caulk or anything like that. It wouldn't hurt, but not sure it's needed.

If you surface mount, I'd use the thinnest piece of gasket material you can get. Or you could skip the gasket completely, and on the back side (after mounting to the wood) you could add caulking or silicone to seal the waveguide to the wood. Because of the waveguides contour,it leaves a perfect area to seal it in with caulking.

Also, when everyone was deciding on screw holes and things like that, we all agreed that the least number of screw holes the better........you could always add more if needed. The plastic on the waveguide is thick and pretty darn solid. But if the gasket material was thicker and you torqued the corner screws down pretty hard, I could see the need to drill a hole towards the middle of the waveguide's flange on the top and bottom to pull it in against the baffle a little.

Personally, I'd probably put 2-3 layers of soft electric tape around the flange as a gasket, then seal it more from the back side. Either way would be fine though.

Sounds good thanks. I am referring to Bills design specifically the Delta 12. I was thinking recessed with caulk as adhesive sealer but that might alter FR. Hes already said recess both drivers same amount. This is for phase and time alignment issues. I was just curious if the lowered diffraction of recessed drivers would be appreciable enough to alter the crossover.

Gotcha, i meant diffraction of tweeter off woofer frame, or am I the one confused?

Awesome.

I still vote for the JBL 

Re all the talk about enclosure volume, I don’t see it as that important; IMO the lion’s share of the job is getting the XO region and above correct.

While knowing the box dimensions may allow BSC (Baffle step compensation) to be designed in, this adds cost/complexity and makes the design less than optimum for the many who will build their boxes to their own unique needs.

I’m not following the reasoning here.

Putting the driver at one end gives the lowest freq mode possible, which are hardest to absorb.

Placing the driver closer to the middle raises the freq, and if it’s in the middle the modes will cancel, same as placing a sub mid-wall.

This seems to invite diffraction issue, and I’m not sure I what the benefit of such an extreme measure is, as the vertical coverage lobe is adequate for most situations.

If I was going to do that, I’d go the other way and rear mount the woofer and slide the basket up behind the WG flange.

But this might require a different XO because of the difference in driver offset, though ¾” depth difference only amounts to 20 deg of phase shift.

My experience using Martin King's worksheets tells me every dimension within a box should be chosen carefully. Every dimension will have an inherent node. The combination of nodes that result may provide cancellation or unwanted resonances. (edit in)

In an "optimized" box the bulk of the pass band is ripple free. In the examples below there is one occurrence of ripple centered ~700 Hz.

With the driver close to the top of a flat top box, we may have an overall height of ~28" and a distance for the driver center to the top of the box of ~ 7 3/8". This produces a ripple in the pass band of ~+/- 2.25 dB.

When the top of the box is sloped down 3", front to rear, the effective height of the box becomes 1 1/2" less, 26.5". The distance from the driver center to the top of the box is also reduced by 1 1/2", 5 7/8". This change in box geometry (the only change) nets a reduction in the ripple to ~ +/- 1 dB...half that of the flat top box.

The Fb tuning of both boxes remains essentially the same because the distance from the driver to the port opening has not been changed.



The plot above is for the box under development, not the final box. I hope it serves to illustrate my point.


Edited by bg40403 - 7/3/12 at 6:37am

Does the worksheet account for damping material?

In any case, this highlights an advantage of sealed boxes - you can totally fill them and more or less kill internal reflections.

Yes, the worksheets account for stuffing, also varying the density along the line.

Refuge in sealed boxes is short sighted, though they may be entirely successful...IMHO...that should be a "systems-based" decision.

What line?

We've been talking about vented and sealed boxes.

How is a sealed box "short sighted'?

What the heck is that supposed to mean?