Top contenders in their class 15,18,21 inch drivers - Page 2

No, I haven't. But I understand they're pretty rare. I also haven't used the Lambda TD-series drivers. Had planned on building high-efficiency mains with the TD15 and a BMS concentric compression driver. But then I heard the big Tannoy Dual Concentrics and just lost interest in other mains.

For whatever reason, the last driver I bought from John was a TC-made (3hp motor) HE15, back when AE was still called "Stryke." It may have been a "basslist" special; I don't remember if his website was even up yet. (I believe that was before TC screwed John with a load of defective parts.)

I give that driver credit for teaching me that long xmax is just a recipe for awful sound if it's not combined with techniques to lower inductance, etc. But not for much else. It was amazing to me how much better-sounding a Peerless XLS12 (the original 830500) was than that driver, despite the considerably greater output capability of the HE15.

Still, the AV15 looks like an interesting driver. Not terribly available, but interesting.

In truth, it depends a little on one's goals.

Geddes suggests that might not be the case. At least in a properly-designed system with adequate signal processing.

I'd like to test the theory at some point. I actually have the woofers to do it, The "good" woofers would be three Aura NS12's, the "lesser" woofers would be three ur-Peerless XLS12s (830500's). Of course the XLS12 is only really "lesser" when compared against something like the Aura. I could also buy three Peerless SLS12's, which are also not really "lesser" in quality but are considerably cheaper at ~$70 a piece.

But that takes time and there's a reasonable possibility that I and other people who have spent money on elite woofers might not like the results at all...

First Lord of the Bass and now an actually Lord

Do we now have to take commands and respond "Yes my Leige" in response

Happy belated B-day by the way my Leige!

I'm running two Dayton HF 15's in 5.2 cubes sealed each on O-audio 500w plate amps in the media room. They are not expensive or fancy, but they definitely overshoot reference levels in THX and require substantial trim and low gain to be flat with the rest of my speakers. The room is on the small side though (18x14) so it's to be expected. I really need a receiver with .2 dual independent subwoofer processing; I think the output was more impressive before Audyssey and when I was testing them one at a time.

ds-21, don't read me wrong...we are pretty much on the same page, just debating the particulars.

"How is the MTX driver "quite similar?""

power to weight ratio.

(bl^2/re)/mms...0.31 for the jbl w15gti, 0.28 for the mtx 9500-04.

w1500h is 0.62. lms ultra is an "in between" design at 0.38.

the w15gti and the 9500-04 model quite similarly in tapped horns too.

"That said, it is the woofer Danley used in the Matterhorn"

didn't know that. very interesting. any chance that you or anyone else has hornresp'd the matterhorn? i'd love to see that.

"I'm not sure what you mean by that. Xmax/weight? BL/weight? Please clarify."

(bl^2/re)/mms

motor strength numerator...moving mass denominator.

compliance figures in as well, but i think that is a secondary consideration. a low qes is what i had previously focused on, but that was incomplete. a low qes suggests a good power to weight ratio, but a higher qes does not necessarily mean a low power to weight. it could just be the result of a stiff suspension.

TH50 driver as well. It is a very strong, robust driver. I have no idea what is actually going on in the motor or whether it has any inductance linearity techniques at all, but it is a beast with a very strong field, a heavy 4" vc and a lot of xmech room. Very underrated.

Ah. Thanks for clarifying as I'd never heard the phrase "power-to-weight ratio" used in this context.

But I just don't buy the BL^2/Re number as signifying anything important. I know that TC Sounds has been pimping it as a figure of merit since forever, but it does not seem to correlate to subwoofer performance in my experience. Good drivers are just all over the place.

By contrast, I have found in my own experience that lower normalized inductance (Le/Re) correlates strongly to preference. Woofers such as the Auras and the W15GTi are all very low in inductance.

So BL^2/Re is not something I look at, unlike trick motor designs, shorting rings, etc.

And FWIW no slight intended to the MTX 9500. But from the little I know about it (what's on their website and the pictures of the Matterhorn) it's not really designed for upper bass linearity in the way the others I listed are.

IMO, the two factors that really distinguish "supersubs" from "very good subs" are thermal performance and performance in the 40-160Hz octave pair. Thermal performance, don't expect anyone will disagree there. Upper-bass performance might be more controversal. But below 40Hz they're all just bass pumps. And a woofer that's just a good bass pump is IMO not a "top contender." (Ahem-TC3k-ahem. Ahem-W7-ahem.) And even if one crosses over at 80Hz, 2d order, performance in the octave above the passband is imporant.

BL^2/RE combined with mms is what governs efficiency and control of the moving assembly. very important IMHE. Inductance as a simple low number doesn't mean much in my experience and I've seen plenty of designs touting shorting rings that weren't really anything special (I always look for them though). The value itself doesn't mean a whole lot. It can be high and still be a great driver it is the variation in inductance with the movement of the vc with increasing distance from center position in both directions that is more important. The LeX curve alah Klippel.

May as well do my own list while i'm here. Some are better for the music range, some for HT bass and others are great at both.

B&C 21SW152, 18SW115
18Sound 21NLW9600 and 9000. 18NLW9600 & 9000
Faital 18XL1500
BMS 18n860
TC Sounds PA5100 and LMS Ultra 5400
JBL 2269H
RE XXX 18d2 (Bit of a one trick pony but still bass pump extraordinaire)
Aura NS18
AE TD18H+

#37, buy what you wish. i was just kicking in some info for consideration.

"BL^2/RE combined with mms is what governs efficiency and control of the moving assembly. very important IMHE."

i've been crapped on for the past several years for suggesting that what you say is true. :-( who cares? let's just all learn and move forward! :-)

Agreed, which is why I specified normalized inductance. That is, inductance divided by DCR.


I agree with you on the latter point, but I've never heard a high inductance driver that meets my standards for bass fidelity.

The closest thing to an exception was the Peerless XLS12 (830500), at 4.2mH/3.5Ω according to the spec sheet. But my measurements of my three XLS12's, and all of the other third-party measurements of that woofer (and for that matter every third-party measurement of a Peerless XLS, XXLS, or SLS woofers that I've seen) were much lower than spec. Often on the order of a third the published value. So I assume they're just measuring it differently.

"Inductance as a simple low number doesn't mean much in my experience and I've seen plenty of designs touting shorting rings that weren't really anything special (I always look for them though). The value itself doesn't mean a whole lot. It can be high and still be a great driver it is the variation in inductance with the movement of the vc with increasing distance from center position in both directions that is more important."

not the first time that i have seen this comment, but it is worth reading...footnoting...researching...and bookmarking. it is a really big idea. even mr. mckinney has suggested that getting the inductance linear is 75%+ of the game while the absolute level is the remainder.

ds-21, i like your "gloves off" approach. we learn the most with such. your hard hitting questions draw out the best. keep punching man...

Goes to show that people are considering them as "contending" for very different things. My list is based entirely on performance reproducing music. I don't care about special effects performance. At all.

The 2269H you list is just fantastic, on paper at least: split-opposed 4" coils, 19mm xmax, and only 21lbs. And JBL drivers tend to live up to their paper performance and then some. (Ditto B&C drivers. Never seen a dud from them.) Hell, the old 8" JBL woofers I'm using right now as car midbasses (116H-1's) are fully competitive with the Peerless and Seas Lotus 8's I've used. ("Lotus" is the car-fi variant of Seas' Excel line.) It really seems like the Scandinavians just learned in the 1990s what JBL's engineers knew 15 years prior.

If I ever buy another subwoofer driver (a doubtful proposition at this point, admittedly) it will probably be that driver.

Another observation: few of these drivers, except some of the pro ones, are that new. Most of them could have been purchased in 2005. Or at least versions of them could. (One couldn't get an Aura-knockoff TC5100, for instance, but one could get a real Aura 1808/Seismic 6196.)

I don't have any reason to doubt that. It makes intuitive sense, too, as really inductance should only set the HF corner of the driver. And the Klippel measurements of drivers I've liked show very linear Le(x) graphs.

But, when Le is lower there's a smaller band in which it can vary.

Also, low inductance could very well be not very important in and of itself, but just something that happens when a driver is designed at the highest level.

Good thread.

Subscribed.

ds-21, that is where i am coming from. ;-) ours are shared thoughts on driver design. maybe you could spec out the common ground where we agree and the areas where we disagree?? then, i if i have anything to contribute, it may have some usefullness. otherwise, i feel like i am just pissing in the wind. not to you, but to everyone else.

"TH50 driver as well."

thanks. but the basket and subsequent modelling gave it away.

that drive actually has some nice features, such as the stiched surround/cone.

I realize that an IB sub is essentially an infinitely large sealed design. Tell me, does your "Top Contender" recommendations stand for an IB implementation?

In each size category, which would be the best performer,...regardless of cost?

What I've read, everyone points toward the AE IB15", and the Fi IB318", however for the 21" category, I don't know, of the current offerings, which would be best. Any thoughts?

Thank you

Interesting way of looking at it.

I think I still prefer Qe, as high P/W can be sabotaged by excessive suspension stiffness/high Q.

This wouldn't matter much in a small box/horn chamber where the total stiffness id dominated by the air spring.

"I think I still prefer Qe, as high P/W can be sabotaged by excessive suspension stiffness/high Q."

noah, you and i are in the same boat. i suspect that we would both start with a strong motor relative to moving mass, then turn up the suspension stiffness for the application. the theoretical ideal case would be a motor that is unlimited in strength, a moving mass of zero, and a suspension with infinite compliance. the problem with that is one ounce of excess power and the driver blows up. this may be one area where the "pro's" have a leg up over diy, as they can build limiters into their works by sacrificing a few drivers in the process...not something the average diy'r would be game for.

I agree that our area of agreement is broad, and our area disagreements narrow. So posting the things we agree on would add a lot to what's going to be a very long post. The three main disagreements (assuming the third isn’t just semantics) I see are as follows:

1) What is the predictive power of your power-to-weight calculation (BL^2/Re/Mms)?

I hadn’t figured it for many of the drivers I’ve mentioned, but here it is (I cheated and used your numbers for the W15GTi):

For reference, the BL^2/Re/Mms index for the other drivers I mentioned. (You already listed the W15GTi and TC LMS-Ultra)

Aura NS-series:
NS10-794-4a: 0.46 g/Ω (Qes = .42, 80mm vc)
NS12-794-4a: 0.37 g/Ω (Qes = .43, 80mm vc)
NS15-992-4a: 0.31 g/Ω (Qes = .37, 100mm vc)
NS18-992-4a: 0.23 g/Ω (Qes = .5, 100mm vc)

JBL W15GTi: 0.31 g/Ω (Qes = .51, 2 vertically split 76mm vc’s)

TC Sounds
TC2+ 15 (Oaudio OEM): .28 g/Ω (Qes = .4)
LMS-Ultra: .38 g/Ω (Qes .34, 100mm vc)

Exodus Maelstrom-X: 0.25 g/Ω (Qes = .43)

Peerless 830500: 0.53 g/Ω (Qes = .21, 51mm vc)

Dayton Reference 15HF: 0.23 g/Ω (Qes = .49, 64mm vc)

All that number really says to me, based on my experience with those drivers, is, “how much EQ am I going to need to extend the response down low?” Which makes it not so different from Qes.

2) What is the predictive power of normalized inductance (Le/Re)?

I say that low normalized inductance (well below 1mH of inductance per ohm of resistance) seems to be from my experience to be a necessary, but not sufficient, condition for a great-sounding subwoofer driver. Others seem to disagree, or place less emphasis on it. I readily concede that low normalized Le could simply be an indicator of best practices design, and not something necessarily targeted in driver design.

3, maybe) Can a “car audio” driver be considered “elite?”

I say yes, based on the same criteria used to judge other drivers. Those include first subjective performance; measured performance such as BL(x), Le(x), and Km(x); and thermal properties (ability to measure similarly after being slammed as when cool). As well as “paper racing” qualities such as “does this subwoofer incorporate some unique trick design in its motor?” I say the JBL W15GTi qualifies, primarily because of its outstanding thermal performance.

Frankly, for all we know JBL might not be using the W15GTi’s platform more widely for home use because of cost and packaging reasons. (They did use a very similar driver in the original Vertec-series sub, but I think it’s safe to assume they’re not using it more widely for pro use because it’s so heavy.) Certainly, it looks like a more expensive driver to make than the old 15” JBL drivers in the ur-Revel Ultima sub. There’s a lot more metal there. It’s also considerably deeper. Also, I wonder if it’s somewhat snobbery-related. (“Oh, the car guys picked up that pro driver, so it must not be worth bothering with for real audio.) It could also be true, of course, that the drivers they make but won’t sell to us are legitimately better. In which case, I just have one thing to say: damn! The W15GTi is already such a great performer, that the other ones must be really, really, really special.

Ds21,

Call me crazy but I've never been really blown away by the w15gti. It's klippel data didn't look exceptional and while it does have a lot of tech and hype it doesn't really separate itself in any of the measurements I've seen. I understand it's dd with braking and shorting paths and all of that, but I think that I would rather have an Lmsr15 personally. Im confident that the pro jbl 2269h is easily in another league of performance with a larger 4" motor, roughly the same xmax, more sd, more motor force, higher efficiency and all while weighing less. If there is a product line where they performance of the drivers must be maximized as much as possible it would be their pro audio line. You can get that driver if you really want it but it is not cheap. I don't know maybe all of the hype around the w15gti is clouding the issue for me. I haven't personally used the driver either.

The bl^2/RE/MMS/cms relationship governs efficiency and perhaps ringing or overshoot as well? Thinking out loud here...Its not just the small signal at rest motor force, impedance and cms that should be considered but all of them in relation to each other plus the mms as they change at each point throughout the drivers useable range of stroke. A sort of a complex calculated summary curve. Perhaps one driver is more powerful near rest but drops off precipitously after 15mm p2p while the technically weaker motored one may maintain much more linear operation to 60mm of travel. Which is the stronger driver? Depends on the application.

"1) What is the predictive power of your power-to-weight calculation (BL^2/Re/Mms)?"

that is based mostly on hundreds of subjective observations. i defer to such vs. my own.

ricci's comments in #55 are spot on.

"3, maybe) Can a “car audio” driver be considered “elite?”"

yes.

Have you played with one? It's the driver that single-handedly made me rethink my thoughts on Linkwitz Transforms.


I wouldn't. The inductance is just too high.

But then again, I'd never consider the XXX 18 for, well, anything.


You'll get no disagreement from me that on paper a 2269h should be better than a W15GTi in every way, because of the bigger coil motor of the same design, and greater volume displacement. But I've never used one. I would like to, though, now that I've heard of it. Dammit!

Except on the efficiency issue. In most of a subwoofer's range, box size, not the driver, determines that.


Indeed. The rumor mill has it that the W15GTi is based off of the platform of the original subwoofer driver in JBL's flagship Vertec line, the 2256G.


I've never heard any hype about the W15GTi, to be honest. Never read a review of it, or anything like that. I'd bet that most of the posts about it on this forum about it are from me.

I will say one negative thing about it, though: WTF have they been thinking about the dustcaps? They're all ugly.

I think you are mistaking our original pro motor (which was an Aura knockoff from the motor standpoint) for the 5100? Our older Pro motor often called the bhive was available in 2005.

Ricci posted about the driver relationships that are of greatest importance.

Away back when I had hair I did some serious driver design. Mechanical linearity has been hinted at as being important. It is in that the distortion products are directly related to how well the motor system can keep the cone system in check. No amount of Faraday rings or pole caps will take care of a non-linear mechanical system. Those refinements are for keeping the midbass and low midrange clean and clear when the rest of the driver has already been optimized.

The most critical after the moving system is taken into account is the motors BL linearity. When you get access to true driver measurements like a Klippel report you usually fall off your chair laughing. Then you get up mad at the marketing pinheads that will massage the numbers to make things look good.

Some of the most posted drivers in this thread are the Aurasound stuff. There is a bit of smoke a mirrors in the specs even with this driver. Dynaudio was famous for this. As are most of the car audio companies listing X-mech as X-max. What has been touted as 75mm of X-max looks great. But what you are getting at that level of excursion is not within the realms of truth and clarity. They can move in and out a total of 75mm mechanical. But they are no longer in the control of the motor. The distortion products go way up past 10%. The BL product is the number of turns of the voice coil that are still saturated by the same amount of magnetic flux density as the when the driver is at rest. Every driver will have a smaller BL product as the cone is moved forward or backward. The question is how far can it move before things poop out. Aurasound uses a good design to get a stable BL product. Most drivers that use a divided gap technique like X-BL or some derivative have a stable BL product over a greater amount of excursion.

What to look for in the graphics we see on driver vendor sites are extended pole pieces, top plates in two sections. Really thick top plates and short coils are another way to get a long linear BL. The better designs use the minimal amount of metal in a saturated field like the Aurasound beehive motor. These optimized motors cost lots to design and lots to manufacture. Either CNC machining or very custom forging dies. Nothing is free!

To get a high flux density in the smallest package and lightest weight some vendors offer the driver with a Neodymium magnet structure. But the same effects can be had with different kinds of magnets. The trade off is the cost of the magnet. Ferrites are almost always cheaper.

The cost of a driver vs it's performance is one of the greatest equations to take into consideration. I haven't seen any posts taking about that.

One more bit oh gripe from a design stand point. Pouring power into a driver becomes a none issue very quickly after 500 watts input. Your largest gain in SPL is at 128 watts. That gets you a 21 dbs of gain over the one watt rating. Double that and you get 256 to get another 3 db. Double that to 512 to get another 3 db etc.

3 db is the amount that most people call perceptually louder. 6 db is noticeably louder and 10 db is twice as loud.

Quickest and biggest bang for the buck is to take a second driver in another enclosure and drive it with the same amp. 3db gain right off the bat. If your amp can handle a parallel load you can get 6db over part of the passband. That's like taking a 250 watt sub amp and turning it into a kilowatt sub amp. Because effectively you have raised the power efficiency by 6 db.

You also cut your distortion products in half.

Mark