Linear motor technologies. Are they the future?

[Looneybomber]

With the introduction of XBL^2 with Adire Audio and LMT/LMS with TC-Sounds, there's been a lot of new products with a linear motor technology, something that's been researched since before most of us were born.

For those that don't know, a linear motor technology, regardless the method, is a means of making a driver's cone move with a more linear amount of force propelling it. Think of it as a nice long flat torque curve. This idea is catching on so well, we not only have subwoofers with XBL^2 motors, but also midranges and even a tweeter!

The ultimate goal is a better performing driver, with less distortion, less inductance, and less power compression...among others. The cons however is it's more expensive to implement in relation to a classic overhung motor (which so many boring speakers are designed from including Kicker, MTX, Fosgate, JL ect...) and static measurments do not look as nice on paper as what the driver actually performs, because dynamically is where the LMT actually beings to work.

So initially you see a driver that costs more to build - and thus to buy, yet doesn't look so hot on paper. However it will outperform a like driver with an (classic 1950's design) overhung motor.

Thoughts?

[Looneybomber]

An excellent post by Dan Wiggins (a transducer engineer at Adire Audio and one of the inventors, if not THE inventor of patented XBL^2 technology) talking to Kyle Lee, (another transducer engineer from TC-Sounds and one of the inventors of patented LMT/LMS) about the XBL^2 topology.

Kyle,

I think you're looking to standardize too much. Amplifiers are really voltage devices, so you want the impedance to be equalized; don't look for even power, look for even current flow (since force is BLi, you want the i to be the same). Which means, if your input is in terms of power, you want the impedance of the voice coil to be the same.

This means you can go with more layers, or different gauge wire. And it's still VERY comparable because you want the effective force to be equalized (that i part of the term). It's perfectly reasonable to do so.

Here's a few graphs for a real-world, shipping model product I did. We'll start with the overhung. I can't give the actual Klippel tests, nor identify the model (NDA), but I can tell you that my simulation of the overhung was close enough to what they actually Klippeled that it led the company to try the XBL version.

Anywhere, here's what the overhung had for a BL curve:



This was with a DCR of 6.1 Ohms.

Now we did an XBL model, with the SAME pole piece, a shorter magnet stack, a thicker and grooved top plate, and a rewound voice coil. Overall motor height and weight was the same, and we used the same ID and OD for the magnet (just a shorter magnet overall).



This was with a DCR of 5.9 Ohms.

Essentially, we doubled stroke, and increased BL everywhere by at least 10%. More efficiency from this increased BL; we kept the Mms the same since the customer required a fairly consistent Fs, and didn't want to change the suspension at all (it's really well done).

And I want to reiterate that we Klippeled both speakers, and found the model to be within a 1-2% of either Klippel result at all positions. And the XBL version is now shipping - it replaced the overhung unit in an existing speaker design. In other words, the client tried it, and liked it enough to make the change (and did so gladly).

So let's look at what these two curves mean. Let's start with THD. First the overhung curve, then the XBL curve:



No contest. The XBL motor breaks 2% THD at 8mm one way; the overhung crosses that at 2.8mm. At 10% THD, we find the XBL motor at 10mm, the overhung at 5mm.

How about driver efficiency? Overhung first, XBL second:



Again, not even close. Not only do we have more SPL everywhere, but we have no change in SPL until we're 9mm one way with the XBL; we see a 3 dB loss in efficiency (because of loss of BL) at 11mm one way. Consistent and flat. The overhung? You're 3 dB down at 5mm, and have measurable loss in a matter of a few mm.

This is a SERIOUS issue. This is compression, pure and simple. The average SPL you hear is the average efficiency over stroke. As stroke increases, the average SPL of the overhung continues to drop, and does so rapidly. This effect is classic compression, and is in fact highly audible, and what most people associate with audible compression!

Thermal compression is a non-issue; your voice coil heats in a matter of milliseconds, and takes seconds or minutes to cool down (fast attack, slow release).

BL compression is instantaneous and completely displacement - not power - based. You can't get around it by simply turning the volume up a notch (like you can with thermal compression in low crest factor situations like prosound). It's always there.

What about box alignments? Guys here sweat tiny little changes in box models, but in reality it's a waste UNLESS they're using a flat BL driver. Here's why - let's look at the Qes (which really dominates Qts) of the overhung and the XBL drivers:



What does this mean? As you move, the Qts of the driver changes. And that means the box alignment changes. In fact, the further you push it, the more the overhung starts to suck. Now you can fit the middle, and make it sound OK at mid-volumes, but then it will be overdamped at low volumes. Or if you design it to sound good with the published T/S parameters (low level), the box will get boomy at higher excursions, because now it's underdamped (Qts is up).

The XBL driver will maintain a nice and consistent Qts over stroke, meaning the alignment doesn't change. Design it to sound boomy or tight, it'll stay that way well over operational characteristics.

Anyway, these effects - THD, compression, Qts/alignment stability - are completely discounted with most packages, T/S measurements, and basic analysis usually done. Yet these are the ones that are most audible! Too many people overlook them because they don't understand, don't know they exist, or don't believe it because no one has ever done this before (of course, linear BL drivers hadn't been done until 2001 with XBL).

To wrap up this lengthy post... I've had lots of real-world, independently measured experiences and cases where I could get more BL AND more efficiency AND more stroke with XBL versus an overhung, you need to optimize each driver rather than trying to hold things constant (it doesn't apply), and I put a lot more faith into the audible issues, rather than the static/low-level measurements which apply to only low-level situations.

[the converted]

So is this anything like servo subs of a decade ago? Seems quite interesting though. What kind of dimensions are we looking at for a tweeter though? I'm assuming a mounting depth of .5" isn't an option here.

[Looneybomber]

Well servo subs are a means of controlling the power to the sub in order to force the cone to move different than it normally would. This is good because as you can see by the overhung motor graphs, the distortion is normally much higher. The servo can then monitor the woofers movement in relation to the input and reduce that distortion.

The cool thing is a servo can still be used in combination with an LMT. However, implementations remain the same with designing the driver(s) and adjusting their zero degree phase shift. Unfortunately, what would work great for a car, will not work great for the house, because above (or below depending on setup and with ported vs sealed since ported has two phase shifts) that zero deg. shift, the servo loses control of the sub and all benefits are then lost. A house will require a much greater range and thus requires a much more delayed phase shift...plus a house will require greater excursion due to the more air volume needing to be compressed.

XBL^2 tweeters for excursion purposes are a joke, but that's not why they have them. One of the benefits of the XBL^2 topology is the use of shorting rings that drastically reduce inductance. However, reducing inductance too much, to near zero, causes a rising effect, which, may be useful depending on implementation. For example, if someone wanted to use a waveguide or horn to load their tweeter which helps raise up the bottom end. A rising top end along with a rising bottom could net a flat response with very high sensitivity and reduced distortion.