Gosh, sorry I'm so late to the party. I just now found this thread. Jake had told me he was going to post, but I didn't expect it so soon.
Okay, the Swarm that Jake has is probably the last set that will use a plate amp. By the time that little wooden amp enclosure is added in, it's not very cost-effective. I have since found a rack-mount amp that is more powerful than I need, but it has the nice added feature of a single band of parametric EQ. This is to address any remaining areas of over-emphasis, because as Young Ho noted the ideal setup would have one of the subs raised up near the ceiling to help smooth the floor-to-ceiling mode but in practice that's not really practical.
That isn't really a stereo plate amp - I just set it up to have two sets of outputs. It's actually a mono signal that's driving the four subs. However, my understanding is that very, very few recordings have stereo information below 80 Hz. I am friends with a researcher who once was tasked with finding out how many CDs have stereo bass, so he had all his co-workers bring in their CD collections and he went through them. He told me that he did not find a single one with stereo information below 80 Hz. Now that's not to say it never exists, but evidently it is rare. For those who want stereo bass, the regular production Swarm can be set up that way just by adding a second rack-mount amp.
The holdup on the rack-mount amp version is that it doesn't have speaker-level inputs. I'm investigating several possible solutions to that, but until I get it resolved it really isn't a "complete" product.
For those of you with DIY in your blood, if you're gonna do a SwarmClone you need to get a plate amp or rack-mount amp that has a 4th order low-pass filter, so that if you scatter them widely they don't betray their location.
The theoretically preferred arrangement is with subs widely scattered, as youngho says. Since vertical scattering is impractical, I set them up so that two of the subs have the woofer high and the port low on the "business" side, and the other two have the woofer low and port high. I'm not sure this makes a significant difference, but it makes me feel like at least I tried.
My source for the asymmetrical scattered multisub concept is indeed Earl Geddes. He and I were riding in my car and he described it, and a lightbulb went off in my head. For years I had been dinking around off and on with various "Maggie-friendly" subwoofer concepts. I tried dipole, transmission line, lowQ sealed box, but never really got a good blend.
Todd Welti, author of that Harmon paper, is probably the leading researcher on the subject of multiple subwoofers. His paper looks at symmetrical rather than asymmetrical placement. Rather than get into a shoot-out with his followers over which is best, I'll just harp on the main point that more is better in either case, and you can buy my product (or do your own version) and try it both ways and make up your own mind.
Now nobody asked yet, so I'll ask it: Why a ported box? "Everybody knows" ported boxes are boomy, sealed boxes are better, transmission lines are better still, etc.
Well for one thing, to get anywhere near comparable output and extension from a sealed or transmission line box, the costs would have gone way, way up. We'd be looking at close to three grand, intead of $1700 (less 10% from my ongoing sale).
Second, the ideal shape for a subwoofer's frequency response curve is very difficult to realize with a sealed or transmission line system without equalization, but it's not that hard with a vented box. And what is that ideal frequency response curve? Well, let's look at a "typical room gain" curve and see what it tells us. Here's one:
http://www.speakerbuilding.com/content/1020/rge.gifOkay, what you are seeing is a roughly 3 dB per octave rise below 100 Hz. Now this won't be the case in every room, and might not be the exact case in any room, but since this is approximately what will happen to the subwoofer's output we might as well take it into account. Failure to take room gain into account results in too much energy at the bottom end of the spectrum, making the system sounds heavy and slow.
So, the Swarm modules have a response that falls by approximately 3 dB per octave below 100 Hz, down to about 25 Hz, below which the response falls off fairly rapidly.
Now in the prototype stage I did a shootout with a Qtc = .5 sealed box that I built, one-on-one, subo a subo. The sealed box had a bit more impact on kick-drum, but the tone of other bass instruments was more natural-sounding with the vented box. So, it's a bit of a tradeoff. Except, to have equal out level capability (and still inferior extension), the sealed box requires a woofer that costs over four times as much. Guess who that cost would get passed on to?
So the Ultimate Swarm would probably use equalized sealed boxes, but it would probably be so expensive that you might as well go see if you JL Audio or REL dealer will give you a sweetheart deal if you buy four of them.
Well I kinda got to rambling there a bit. Okay, back to your questions:
Rajacat asked, "Do all the subs have to be of exactly the same construction, size and type?"
The answer is no, and in fact Geddes says that it's best if they're all somewhat different. I experimented with tuning each of the four differently vs tuning them all the same (as low as is practical), and I preferred them all the same by a small margin - in this case, they sounded a little bit tighter to me with the low tuning. But if you already have one or two subs and don't feel like selling them and starting over, you can add others subs that are smaller and cheaper and probably end up in about the same place (and maybe even a bit better). Only thing is, remember that if you're going to scatter them, whether a la Geddes or a la Welti, you want a 4th order low-pass filter rolling off the subs on top, so that you can't hear their locations. I suppose the two near the main speakers could have a 2nd order filter, but any that will be placed far away from the main speakers should have a 4th order filter.
The anticipated maximum output level of the Swarm system is roughly equal to a pair of Maggies driven by a 400 watt per channel amp, assuming the regular production big amp (which is actually powerful enough to cook the little woofers, so a little bit of sanity is called for). With the smaller amp that Jake has, the Swarm can keep up with about 200 watts per channel on a pair of Maggies.
Feel free to fire away with any questions or challenges or whatever. I'll post about unorthodox polarity options another time.
Thank you all for participating here. Thanks for starting this thread Jake, and youngho it's good to see you again.
Duke