LMS-5400 buy-in

[WillyD]

It would show up in their FR then.

Oh but it does show up in a measurement. Here are two near field measurements. One of an AVA15 and the other of the 390. The difference is obvious.

As for high end rolloff on the Avalanche 18, there is some, but it shouldn't cause any dips in response as long as you don't use a crossover higher than 80hz

I won't talk about dips, but since the crossover to the sub isn't a brickwall, the linearity of the driver beyond the xover point is indeed relevent. So it only makes sense that a more linear driver in the upper frequencies would sound punchier than one that rolls off the high end to a degree.

Now yeah, the rolloff in the last graph you posted is certainly not bad. Only ~4dB down at 100Hz, right? Still there though...so I guess all I am saying is...blaming others thoughts about the Ava's lack of punch on their penchant for higher distortion may not be entirely correct. It could be...but I think the ava's roll off also may be a good portion of it. We've yet to be given a reason to think the Ava has abnormally good upper frequency response.

If you were in the states I would probably be up for it, but I wouldn't want to take the risk of one getting damaged (or lost) sending it overseas - you can't really go out and buy a replacement.

Hey, what about a Klippel? I am pretty sure npdang (he's in CA) would test an Ava 18 for free. That'd be much cheaper, easier, and safer than sending one overseas. Quite informative too!

[SteveCallas]

Willy wrote:

Oh but it does show up in a measurement

That's only the sub. If there were a lack of punch due to a dip in FR in the upper bass range caused by the rolloff of the sub above the crossover frequency as you mentioned in your previous post, you'd need to measure up to ~200hz with the mains in play to see it. Seeing as most processors use a 4th order crossover slope, that would put the sub 24db down by 160hz. A couple db difference in output form the sub that high shouldn't really matter, especially when you consider that room effects play a huge role in that range.

Now that said, again, I have't observed such dropoff in upper bass myself, and I know my VR3s don't have humped bass above 100hz which could be making up for it. My response up to 200hz was extremely flat in the old place - haven't measured it at this new place.

Also, I'm not necessarily "blaming" others' perceived lack of punch on low distortion, as I just remembered after your post that they both have pretty bad in room FRs.

[brucek]

LINEAR response graph...

[SteveCallas]

In more ways then one

[WillyD]

That's only the sub. If there were a lack of punch due to a dip in FR in the upper bass range caused by the rolloff of the sub above the crossover frequency as you mentioned in your previous post, you'd need to measure up to ~200hz with the mains in play to see it.

Yeah, I understand what you're saying.

But the fact that those measurements I linked to are only of the sub driver doesn't negate them, if anything they are more useful. Their deficiencies can't be masked. If a roll off exists, it will show up.

Also, I'm not necessarily "blaming" others' perceived lack of punch on low distortion

Well, you kind of hinted at it with this:

"It would be nice to see an Avalanche 18 measured because some owners are saying there isn't enough punch in the higher frequencies. I have to think that be because it is clean near the top, and they want that synthetic punch."

You say it is clean near the top but how clean, relative to what, and where is the evidence?

Seeing as most processors use a 4th order crossover slope, that would put the sub 24db down by 160hz. A couple db difference in output form the sub that high shouldn't really matter, especially when you consider that room effects play a huge role in that range.

Now that said, again, I have't observed such dropoff in upper bass myself, and I know my VR3s don't have humped bass above 100hz which could be making up for it. My response up to 200hz was extremely flat in the old place - haven't measured it at this new place.

But you do use a 60Hz xover, so it'd only make sense that you shouldn't suffer from such problem. Have a FR of just your VR3s? Boundary gain could definitely increase their response around 100Hz.

Just things to think about.

[SteveCallas]

That measurement was with an 80hz crossover, I only switched to 60hz after moving to this new place. FR of the VR3s is the purple line.



I've measured them to be flat to 30hz in room.

Boundary gain could definitely increase their response around 100Hz.

The sub would be the one getting the most boundary gain, as it's in a corner. The baffles of the mains are 3.5' out away from the front wall and the nearest side wall to one of them is ~6' away.

You say it is clean near the top but how clean, relative to what, and where is the evidence?

The inductance isn't bad and it's an 18" driver, so logically it should be pretty good. We can see from Ilkka's measurements that more linearity up top equates to lower distortion, and that an 18" is more capable than a 15" (no suprise). That some people claim it lacks punch would also point me to that conclusion, but again, the overwhelming factor in that would be their in room FR. Relative to what would be relative to whatever they were using as a subwoofer before.

[Chris Brunhaver]

A few comments:

Ilkka is right on the money about the TC woofers having a more linear compliance (because of their 10" spiders with deep rolls) and this having an effect on the distortion measurement below resonance. However, if you have a spider that is more linear in one direction than another (which is many times the case) and you mirror image them, while things can be a little more symmetrical, you kind of get the worse of both rather than the best. I think this was Dan’s main point along with the idea that it doesn’t help rocking as much as is though but that may be debatable.

Obviously the LMS 4" diameter >3" long coil with an average of >4 layers will have higher self inductance (and issues with modulation distortion etc) than a 3" diameter ~1.5" long 4 layer coil like an SDX15 or Tumult. What you're seeing in the 5400 is a difference in implementation of shorting rings etc., not proof that there isn’t an issue with this with the LMS design.

The tough thing about shorting rings is that hey are not effective at low frequency without being really thick. Also, if you sleeve the pole (TC mentions using a 4.5" long copper sleeve) you're making an already very wide gap even wider (since copper has about the same DC permeability as air), further reducing the flux density which makes low frequency modulation distortion even worse. In short, it looks like TC was only able to get good performance with that huge motor structure (4 ~11" OD magnets and 4" coil) to get the Q back down to reasonable. The Eclipe 8200 has shorting rings are reasonable performance but isn’t really well suited to its application (Mms too high, BL too low). And really needs the larger magnet etc. Basically, we're seeing more of an issue with differences in implementation and so you really cannot make accurate blanket statements about the technology.

With XBL^2, the rebate between the top plates is a natural place for a shorting ring (and a fairly thick one can be put there) without lowering flux density in the gap and adding a lot of cost. Ideally, additional shorting rings or a sleeve on the pole need to be implemented to balance things throughout the stroke of the driver.

However, since you have two gaps, it requires a lot of magnet to get the BL high enough and so you see a lot of 4 layer coils and such is all of the subs. Of course the best way to address the inductance issue is to have most of your BL coming form B and with a huge magnet system like the 4 HP (or some big Neo thing) you I'd be curious to see an XBL^2 driver with a longer 1-2 layer coil take on a super woofer.

The inductance graph shown by Dan is a simulation, not a measurement. However, you can see the effect of the shorting ring near the rest of the driver (i.e. between the top plates) and how it is effective more at higher frequency than lower. Of course, inductance increases inward into the motor because of the presence of more steel (making the coil essentially iron core instead of air core). In reality, things mighty be a hair better than this measurements because of some/most of the steel being in saturation, as this sim isn’t taking into account the relative DC permeability.

The Klippel inductance result isn’t at all directly comparable. It is a lumped model which is somewhat skewed by the low frequency side of things (required to get the high displacement). In short, if you make this graph flat, you could be putting too much copper in the motor (making the curve very non-flat in the range where the copper is effective). This is the case with some drivers like the Eminence Magnum and probably also some of the PE stuff.

The Ascendant Avalanche uses the exact sample coil and gap dims as a Brahma. They used a softer spider and use a slightly shorter different magnet (short magnet with a large bumped back plate) but I’d expect the distortion performance to be basically identical (i.e. no revelation with the AVA15 tests). It’s a round wire coil that ~20% shorter than a Tumult and inductance, Max and power handling is a bit lower but they are really more similar than different.

Basically, yhere is a lot more to the total picture and root causes than most understand (including myself) but some guys here know enough “to be dangerous” and campaign in a certain direction that is somewhat misleading.

[Ilkka]

Chris Brunhaver wrote:

A few comments:

Hi Chris,

Hopefully you don't mind me mentioning that you used to work for Adire Audio, so that people know where we're coming from.

As you have probably noticed, I'm more of facts and objective data than marketing talk type of guy. I have no business interests towards any manufacturer making complete HT subwoofers or just drivers for them/car audio. Therefore I also do not campaign in any certain direction, other than towards good performance. Whether that road leads to TC Sounds, JL Audio, SVS, CSS, or any other company, doesn't really matter to me at all.

Obviously the LMS 4" diameter >3" long coil with an average of >4 layers will have higher self inductance (and issues with modulation distortion etc) than a 3" diameter ~1.5" long 4 layer coil like an SDX15 or Tumult. What you're seeing in the 5400 is a difference in implementation of shorting rings etc., not proof that there isn’t an issue with this with the LMS design.

I don't think anyone tried to even hint that it would have lower inductance than, say, an XBL^2 coil. But the way they have handled and tamed it is the impressive thing. For example the MK1 Tumult clearly shows that low self inductance doesn't really count unless also other improvements (shorting ring etc.) are being applied properly.

Measurements show extremely low THD and very flat upper-end frequency response for the LMS coil. In the end that is much more important than just saying that the other design must have lower inductance due to shorter VC.

The inductance graph shown by Dan is a simulation, not a measurement. However, you can see the effect of the shorting ring near the rest of the driver (i.e. between the top plates) and how it is effective more at higher frequency than lower. Of course, inductance increases inward into the motor because of the presence of more steel (making the coil essentially iron core instead of air core).

Thanks for the correction. I accidentally had the inward/forward stroke backwards.

The Klippel inductance result isn’t at all directly comparable. It is a lumped model which is somewhat skewed by the low frequency side of things (required to get the high displacement). In short, if you make this graph flat, you could be putting too much copper in the motor (making the curve very non-flat in the range where the copper is effective).

Ok. I was a bit hesitant to post it because I wasn't sure how comparable they would be. I edited my post to reflect this new info.

[SteveCallas]

but some guys here know enough “to be dangerous” and campaign in a certain direction that is somewhat misleading.

Out of curiousity, what direction has been misleading in this thread? The fact that companies don't share any REAL data is what creates these types of situations, and we are all in debt to Ilkka for actually going out and performing these tests. XBL^2 may have several theoretical and potential advantages, but Ilkka's testing of the Tumult and LMS goes to show that theoretical is one thing and real world is another. As an Avalanche 18 owner, I'm happy knowing its performance lies somewhere between the two for a fraction of the cost of either

[WillyD]

The inductance isn't bad and it's an 18" driver, so logically it should be pretty good. We can see from Ilkka's measurements that more linearity up top equates to lower distortion, and that an 18" is more capable than a 15" (no suprise). That some people claim it lacks punch would also point me to that conclusion, but again, the overwhelming factor in that would be their in room FR. Relative to what would be relative to whatever they were using as a subwoofer before.

Someone just needs to send an Ava to (npdang ) for a Klippel. Its long overdue. Seems like more than a few people would be willing to chip in too.

[Chris Brunhaver]

I did mention on the other thread that I did work for Adire and don't mind you mentioning it. Of course I haven't worked there is a couple of years but still am working the the speaker business at another company. However, I would think that this would help m credibility not hurt it, as I have had a lot of experience with these sots of drivers.

I'm not a hardcore engineering type and know more than some but an still in the "enough to be dangerous" category as well. I just think it is a mistake to come to conclusions about a particular technology or approach like LMS or XBL^2 from the particulars of a given driver and many times the dominant factors are the particular execution and trade-offs made.

A avalanche is essentially a Brahma and so isn't between a Tumult and 5400 at all. Tumult is is its bigger brother with significantly high power handling and stroke and BL but with somewhat higher inductance as well. The 5400 is something different altogether. There are some measurements of a Brahma 12" based ported woofer in Keith Yates "way down deep" article here:
http://www.ultimateavmag.com/feature...ay/index1.html
As you can see, it is similar to the Tumult with less output but also a little less upper end modulation distortion. Still, that inductive hump is there.

Klippel has real problems with a Brahma / AVA as these drivers as they are suspension limited enough that you can't get a proper BL curve fit. I don't think it will tell you what you're looking for and the distortion measurement approach is probably better.

[Ilkka]

Chris Brunhaver wrote:

I'm not a hard core engineering type and know more than some but an still in the "enough to be dangerous" category as well. I just think it is a mistake to come to conclusions about a particular technology or approach like LMS or XBL^2 from the particulars of a given driver and many times the dominant factors are the particular execution and trade-offs made.

So are you saying that neither the AA Tumult nor the CSS SDX15 represent the XBL^2 properly?

If yes, may I enquire which driver does then?

[SteveCallas]

A avalanche is essentially a Brahma and so isn't between a Tumult and 5400 at all.

The Avalanche 18 as compared to a Tumult 15 has a flatter top end, more displacement, and one would have to imagine less distortion at a given output level due to less travel being needed. Power handling is more than adequate for a LLT application, to which it is perfectly suited.

On top of that, when you take the price tag of the Tumult into account, based on Ilkka's tsting, I'd be nice if I said the performance wasn't exactly stellar.

[WillyD]

SteveCallas wrote:

The Avalanche 18 as compared to a Tumult 15 has a flatter top end, more displacement, and one would have to imagine less distortion at a given output level due to less travel being needed. Power handling is more than adequate for a LLT application, to which it is perfectly suited.

On top of that, when you take the price tag of the Tumult into account, based on Ilkka's tsting, I'd be nice if I said the performance wasn't exactly stellar.

Well do remember that the Tumult Ilkka tested was a Mk1. The Mk2 had at least this change:

"New for the Mark II design is an integrated shorting ring, lowering inductance even further and drastically lowering flux modulation. "

But who knows how significant the difference really is?

Attached are some close mic FRs I took of the ported Avalanche 18 using an 80hz and 200hz crossover with all speakers off.

And here is one of my TC-2000.



Doesn't quite match up to Ilkka's measurements does it? Maybe we shouldn't go around posting our measurements taken by RS meters as the absolute truth.

The inductance isn't bad and it's an 18" driver,

Doesn't appear to be any lower than TC-2K's either, yet you've said the TC-2ks inductance is unfortunate and it should be crossed over at 60hz. How is the Ava any different?

[Ricci]

Illka,
Are the enclosure/s for the LMS's done? Do you know what alignments you'll be going with and can you give an estimate of when the outdoor testing will take place? Sorry if you've already stated all of this info before.

[Chris Brunhaver]

So are you saying that neither the AA Tumult nor the CSS SDX15 represent the XBL^2 properly?

The the extent that XBL^2 isn't that well suited for this sort of application then yes. For small box subwoofers you need high BL, high Mms and high power handling and this isn't a great fit for XBL^2. It is better suited to moderately low BL and Mms and wider band drivers. If you jump over into the Fi forum and read Scott's comments (a man with a lot of experience with this stuff), I think that they are right on the money.

All of these use fairly high inductance copper coils and the Brahma / Avalanche could have used shorting rings and we gapped a little wide (though not as bad as a Tumult, which was gapped for an 8 layer flat wire aluminum RE XXX coil). At least the SDX15 has a shorting ring (to address this somewhat) but this could probably be improved somewhat (at a cost). The Mark II Tumults were quite similar to the SDX15 (with a hair bigger magnet) and the same comments apply. I didn't really get a chance to play with these (though a friend has one) but I don't think they had as good a spider or as rugged a build quality as the originals.

The Avalanche 18 was under motored, as is the new stuff from DIYcable. Steve, I'm glad you like your 20+ cubic foot EBS enclosure but I'll pass. This, in part, is why Adire didn't do one (i.e. Tumult of Brahma 18) and also that slow Maelstrom sales had scared them away from 18's.

Probably the best implemented XBL^2 design were the CSS FR125S (though it could use a bit more magnet and better shorting ring implementation) and RE XXX 6 (he original Extremis) which had the an awesome motor design though the mass was a bit high and soft parts ho-hum.

For subs, the Blueprint 01 (w/ XBL^2) / Koda motor was a pretty elegant solution. With some shorting rings and a bigger magnet, this 22 mm top plate ~16 mm Xmax solution was a bit of a sweet spot.

That is of course just my $.02. While they are cool to talk about and get a lot of attention, I'm not really in favor of the single huge subwoofer and would rather do a few smaller subs spaced throughout an HT.

[Ilkka]