Teeny tiny bandpass filler sub with incredible numbers: what am I missing?

[DS-21]

I'm getting some numbers that are pretty incredible for a XLS12 in a tiny, tiny 4th order bandpass box: 12L closed, 6L tuned to ~60Hz. I realize the XLS12 is a unique driver, with a well-optimized suspension, huge motor force (Qes=0.21!), low Fs, but moderate-by-modern-standard stroke. But still, the numbers are so incredible that I, frankly, don't believe them. Here's what Unibox is telling me:

 

 

F3's of 30 and 130Hz, AND on 105dB down to 30Hz on 400W. Near ideal numbers for a "Broadband" sub in a Geddes-style system.

Or, if one wanted to use four of them in a room, one could get a paper 118dB over the passband, and a solid 110dB down to 20Hz. Perhaps not quite supersub territory...but not bad at all for ~80L of cabinet and 1.6kW!

 

(And given the numbers posted by the Genelec HTS6 in several tests, and the SVS PB12-NSB here on data-bass, those numbers don't seem far-fetched to me.)

 

The 4th order BP seems to give it a good efficiency advantage down low. Here's the XLS12 based sub compared to an Aura NS10-794-4A in a 15L closed box, which is what I'm currently using as my high-mounted sub. Also powered by 400W.

 

 

I had filed that a way for a while, but vents lead to ungodly resonances, and PR's light enough to tune a 6L box to 60Hz with big throw are uncommon. Most are too heavy. But the CSS APR12 from Meniscus is actually light enough, and in theory should have enough xmax to do it. A 15" PR a la Geddes would presumably be better, but the room calls for a sub that will fit inside a cubicle on one of those ubiquitous Ikea Expedit bookshelves.)

 

So, what am I missing? Will it rip the PR from its suspension the first time it gets serious power? Suffer from resonances that don't show up in the simplified model that severely limit bandwidth? Something else entirely?

 

I'm headed out of town on vacation for a couple weeks starting Saturday, but I'm tempted to order the parts (except for the XLS12, as I have 5 of those already), commission the box, and screw it all together when I get back home.

[Ricci]

Have you checked to make sure that this is actually buildable? 6 liters to fit the PR and clear maximum theoretical active driver plus PR stroke? That seems to be the challenge. Is it enough of an advantage over a simple 25L sealed XLS build to justify the increased complexity and cost? If you can make it work go for it.

[DS-21]

I think it is practicable. The APR12 is light enough (assuming the specs are accurate) to effect the tuning needed. And the mass is up front, not hanging off the back.

 

I was thinking a 12" x 5" Sonotube for the resonant chamber, with a wooden ring to bring the volume down a little bit. The APR12 also has its mass tube in a "snout" ahead of the cone rather than behind it, so it shouldn't go beyond the ~4" depth of the basket.

 

I don't know if the sealed chamber should be a Sonotube or a box. I'm thinking box, just because if it's too light it could bounce around pretty badly.

 

As for the advantage, it loses a bit above 55Hz, but below that it looks like it has a fairly large advantage (considering the size) to me.

 

[Kyle]

DS, can you post a graph of just the standard bass reflex response in the same external volume for comparisons.

[DS-21]

Not for a couple weeks. I'm in Lahore right now, and because air travel had gotten so annoying - Emirates now even has a one-bag-of-under-7kg carryon limit - I only took my iPhone.

 

I expect, however, broadly similar passband efficiency at the cost of not being able to contribute to room pressurization below cutoff, unlike the bp4.

[deepthoughts]

Watch out for the peaks in the response curve. You probably want to go a little larger or adjust tuning for softer corners. The real inductance vs your model assumptions will also shift the shape of the upper half of the response.

 

This works very well with very good sound quality. Fitting the parts, dialing in the tuning, and making sure the parts are all happy in the pressures created end up being the pitfals which simply require correlation of models and testing of real parts. Sometimes the parts work beautifully, sometimes you'll get a surprise.

 

-Mark

[DS-21]

Considering that the xls12 is good to 400+ Hz as a direct radiator, why would Le be a factor up top?

 

Does the bp loading interact with inductance differently that direct radiation?

[deepthoughts]

Considering that the xls12 is good to 400+ Hz as a direct radiator, why would Le be a factor up top?

 

Does the bp loading interact with inductance differently that direct radiation?

 

I don't believe there is a specific "difference" vs. other designs, but take any design with a real woofer and change the inductance of the woofer by 1/2, x2 or x4 and watch the response changes. Higher inductance, especially the shorted inductance (L2||R2) common with big coils and shorting rings, can create saddles in the response making the measured result look much lumpier than planned. In some cases it actually helps or allows different alignments. When playing with bandpass models, I prefer to think of it as shifting around where you use the efficiency of the driver. As you kill top end efficiency, you can use more down low, or as you narrow the total bandwidth you can get greater efficiency.

 

Fortunately the PR type bandpasses follow the models fairly well; at least as accurately as the parameters you input. While ports can certainly be made to work, any dynamic change in behavior becomes more audible when such compression/expansion occurs at the top end of the subwoofer range vs. very low frequencies common in a typical reflex design. The only pain is determining the real tuning vs. what some nominal Sd, Mms & Cms tell you. Real SPL & impedance measurements and correlation to models can make it a fairly quick detail to check or adjust.

[Ricci]

Eventually I want to make an attempt at an 8th order passive radiator band pass but I'm not sure with what yet. Probably some close out 8 or 10" woofer. I already have a slew of TC VMP's of every size. I'm going to try to keep it as small, cheap and simple as possible since there will surely be a steep learning curve involved and I don't want to waste too much wood. Even 4th orders can sometimes be difficult to nail on the first go around. 6th are a bear to get right from what i gather so 8th should be hellish...I don't have any real point to it other than as a pet project.

 

DS-21 you should give this one a shot. It's small so if it's not right the first time the materials shouldn't cost much. The difficulties will probably be in the details. Parameters slightly off from mfg specs, build slightly different from model, tuning, clearances, resonances unaccounted for etc...

[DS-21]

I don't believe there is a specific "difference" vs. other designs, but take any design with a real woofer and change the inductance of the woofer by 1/2, x2 or x4 and watch the response changes. Higher inductance, especially the shorted inductance (L2||R2) common with big coils and shorting rings, can create saddles in the response making the measured result look much lumpier than planned.

 

Obviously. My point is just that, in real life, the XLS12 as a direct radiator actually measures quite flat up to ~400Hz. No measured midbass hump. (As an aside, every time I've measured the Le of one on FuzzMeasure with my impedance jig, I've gotten a number about 1/3 the number Peerless specs.)

 

Eventually I want to make an attempt at an 8th order passive radiator band pass but I'm not sure with what yet. Probably some close out 8 or 10" woofer. I already have a slew of TC VMP's of every size. I'm going to try to keep it as small, cheap and simple as possible since there will surely be a steep learning curve involved and I don't want to waste too much wood.

 

I wonder if the VMP's might be too heavy for a BP, even with no mass added. The only reason I hadn't yet built this XLS12 design is because I hadn't seen a PR that was light enough and had decent throw. The VMP's, as well as the XLS12-PR's I have, are too heavy to effect the tuning I want in this design. And any other BP4 I've modeled, including a big one (~250L gross volume, F3's of ~16 and ~50Hz) with the Maelstrom-X Mk. I.

 

The CSS APR12 looks like the only one that will work in that design. Geddes uses the flat-piston Dayton one for his, but he goes up to the 15 for a B&C 12TBX100, and a 15 won't work in the space I have.

 

DS-21 you should give this one a shot. It's small so if it's not right the first time the materials shouldn't cost much. The difficulties will probably be in the details. Parameters slightly off from mfg specs, build slightly different from model, tuning, clearances, resonances unaccounted for etc...

 

Now that I'm back in town - it's amazing how much one appreciates things like constant supply of electricity, when one hasn't had it for a while; Pakistan is doing this crazy "load shedding" thing that amounts to 10-14 hrs of electricity per day, with no rhyme or reason as to timing... - it's at the top of my recreation list. (Though since I do have 5 of the woofers and only one in current use, I'm also thinking of just doing a dual-opposed sealed with boatloads of EQ...) I'm not worried about mfg specs, because I can measure drivers. I'd fit the box to the driver in question.