LLT Explained - Discussion Thread

[SteveCallas]

The WHOLE thing under 4', or the tube length under 4'? You'd be hard pressed to get 240 effective liters trying to keep the whole thing under 4'. Figure 4" for the legs and 2.25" for the end caps and base plate and that leaves you with 41.75" of tube. A 6" diameter port that is 32" long will only get to a 17hz tuning, as effective volume is ~210 liters.

Could you manage a 24" tube that is 48" long? Using a 6" port that is 29" long, you could get ~15hz tuning with ~300 effective liters. Add in feet and caps and you're looking at ~55". Feed it 600 watts and you're golden. ~112db peaks in room shouldn't be a problem, and extension should be solid to the low teens.

NOTE - Man, that sonosub program Collo wrote really makes the tube dimensioning a LOT easier. A very valuable tool.

[Exocer]

You're right. That'd be about 4' for the tube itself not taking anything else into account.

Collo's program is nice indeed. I've settled on the 24" tube (WillD made it clear that 22" tube might be hard to find) and going a little over 4' is okay.
That sounds like a perfect plan.

[steve nn]

WillD made it clear that 22" tube might be hard to find) and going a little over 4' is okay.

Hey Ex...I had a dickens of a time securing 22" sono!..I was just about ready to give up when I found out that HD had the muscle to order it for me. Anyway 20" and 24" is easy to find here in the Portland metro area. I guess the 22" is considered a specialty size from what I hear.

So your actually going to do a LLT then? Are you in the mode now?..that being you have started on it? Don't forget to post pics of your room after routering way cool

[Exocer]

I am in the mode alright, but I haven't started routing...frankly I haven't decided yet if im ready to fill this room up with MDF dust again any time soon...I'm definitely more ready to hear what an LLT can do than actually build one happen to still have any spare sono's Stevenn?

[steve nn]

happen to still have any spare sono's Stevenn?

Oh dear...I got plenty of 22" sono including drivers, but you have that part covered. I guess I could do a kit for you minus the driver if your interested? The right packaging costs $$ though.

[Exocer]

I am interested, considering i'd spend over $100 on the sonotube in the first place...we'll handle the rest over PM

[davepete]

Steve,
Excellent write-up on the LLT concept! Very well done.

I'm curious, having taken a look at your sonosub using the Ava18, I'm curious - what are you powering it with? Looking at a model in WinISD of a similar sub using the TC-Sounds LMS-5400 18" driver, it appears it would require a VERY large amp to push this sub to its full capabilities. I'm assuming the Ava18 would also. In fact, it looks to me as though for the LMS it'd require about 6000 watts.

Also, on a somewhat related note, I've noticed what looks to me like a bug in the sonosub dimension calculator you mentioned. In attempting to put some dimensions on a possible sub using this driver, I tried modeling a 30" internal diameter tube. Everything looks fine until I check the 'endcaps' page where it says the diameter of the endcaps should be 23 5/8". Now, using the math I was taught in school, a 23 5/8 inch disc would fall right through a 30" diameter tube. What am I missing here?

[WillyD]

davepete - The LMS-5400 would need or preferably have about a ~2200-2400W amp (like an EP2500) to get the most out of it since it has very high power handling, higher excursion (11mm more than the Ava, even more if you use 70% Bl) and lower sensitivty.

It most definitely doesn't need 6000W though...


The Ava18 only needs around 600-700W.

[SteveCallas]

I'm curious, having taken a look at your sonosub using the Ava18, I'm curious - what are you powering it with? Looking at a model in WinISD of a similar sub using the TC-Sounds LMS-5400 18" driver, it appears it would require a VERY large amp to push this sub to its full capabilities. I'm assuming the Ava18 would also. In fact, it looks to me as though for the LMS it'd require about 6000 watts.

The rated sensitivities may be similar, but in the same size enclosures, the Avalanche 18 will produce more output with less power. With my enclosure, tuning, and desire to stay amp limited, I wanted 600 watts into 4 ohms. I found it in the Carvin hd1800. Realistically, I'm probably rarely pushing more than 50 watts. The LMS with 6000 watts wouldn't fit the amp limited criteria - 2000 watts is all you'd want.

Also, on a somewhat related note, I've noticed what looks to me like a bug in the sonosub dimension calculator you mentioned. In attempting to put some dimensions on a possible sub using this driver, I tried modeling a 30" internal diameter tube. Everything looks fine until I check the 'endcaps' page where it says the diameter of the endcaps should be 23 5/8". Now, using the math I was taught in school, a 23 5/8 inch disc would fall right through a 30" diameter tube. What am I missing here?

Well it's not supposed to calculate that for you - end cap diameter is a matter of taste. Some like it flush, some like a small overhang, some like a large overhang. It defaults to 23 5/8" for the default design Collo has loaded. He included fields like end cap diameter, baseplate diameter, port covers, and other such things just so that you can get the design settled all in one area.

Excellent write-up on the LLT concept! Very well done.

Thanks.

[davepete]

SteveCallas wrote:

The rated sensitivities may be similar, but in the same size enclosures, the Avalanche 18 will produce more output with less power. With my enclosure, tuning, and desire to stay amp limited, I wanted 600 watts into 4 ohms. I found it in the Carvin hd1800. Realistically, I'm probably rarely pushing more than 50 watts. The LMS with 6000 watts wouldn't fit the amp limited criteria - 2000 watts is all you'd want.

Willy, Steve, thanks for the responses. Steve, I'm still trying to understand your "amp-limited" concept as you conceive it. How are you determining amp size for a particular application to make it properly amp limited? Also, I'm trying to understand exactly what the limiting factor would be in this case to keep the power at no more than 2000W. WinISD shows me that at that power level in a 19 cu ft enclosure tuned to 12 hz (using the 18" LMS), we're only using 38mm of the driver's 76mm p-to-p Xmax (above tuning). Clearly the driver's not close to excursion limits at 2000W input. It also is rated to handle much more than this thermally (8000W dynamic). The only issue I can think of is how quickly the driver gets to Xmax below tuning. At 2000W, the model shows the driver hitting excursion limits about 8.8hz. At 3200, it hits at about 9.7. I can't see either of these being a real problem, would you agree? Now, obviously modeling a sub can be somewhat different than real life performance, but where am I going wrong in thinking a sub like this could easily take 3200W? It would seem at that level it's only working the driver at about half its capabilities. It would be pushing an 8" port pretty hard, though, at 56m/s air velocity.

As a quick comparison, TC-Sounds themselves now offer a PR sub design using the 15" LMS 5400 (not tuned or sized like an LLT, though, I believe) which they match with a 3200W amp.

[WillyD]

Hey Dave,

Steve is talking about using an amp that can't put out enough power at a given load that is enough to bottom a driver out above the tuning frequency. Thats all I'll say concerning that..

You're a bit confused bout the xmax and modeling. In WinISD, the xmax is one-way linear, which for the 5400 is 38mm. You don't use the peak to peak number in simulations. However, the xmech of this driver is close to 50mm one-way...

Now, obviously modeling a sub can be somewhat different than real life performance, but where am I going wrong in thinking a sub like this could easily take 3200W? It would seem at that level it's only working the driver at about half its capabilities.

No, it would be working the driver at full capacity.The d

I do happen to think that the LMS-5400 and possibly the upcoming 4100 are better suited to passive radiator subs because they seem to like smaller enclosures which would problematic to port properly. They don't need 19ft^3 to work well, you know? And in a smaller enclosure, they can handle more power (like a full 3200W).

[davepete]

WillyD wrote:

Hey Dave,

Steve is talking about using an amp that can't put out enough power at a given load that is enough to bottom a driver out above the tuning frequency. Thats all I'll say concerning that..

That makes sense. I would agree with going that route in the absence of having some kind of limiter capability.

You're a bit confused bout the xmax and modeling. In WinISD, the xmax is one-way linear, which for the 5400 is 38mm. You don't use the peak to peak number in simulations. However, the xmech of this driver is close to 50mm one-way...

Hmm, it clearly states "peak" right on the input screen. In addition, it appears to use that number as though it was p-to-p when calculating the VD. Let's take an example. Using the numbers from the TC-Sounds website for the LMS 5400 18", the Xmax is 38mm one-way. We've also got a listed diameter of 39cm. That gives an Sd of 1194.6 sq cm if we ignore special handling for the surround. Then multiplying that by twice the one-way Xmax gives 9.08 liters, which is the value WinISD calculates if you assume the Xmax it requires is peak to peak. So, respectfully, I believe you're incorrect about the way WinISD handles Xmax.

No, it would be working the driver at full capacity.The d

I do happen to think that the LMS-5400 and possibly the upcoming 4100 are better suited to passive radiator subs because they seem to like smaller enclosures which would problematic to port properly. They don't need 19ft^3 to work well, you know? And in a smaller enclosure, they can handle more power (like a full 3200W).

Certainly what you wrote is true of TC-Sounds' TC-5200 line of drivers. The 18 incher in that line does not model well for an LLT at all, requiring a much smaller enclosure IIRC, not nearly enough to put in a big port. That driver would have to be tuned in a PR enclosure if one chose not to go sealed with them. Not true of the LMS-5400 18", though. For example, if I drop the enclosure size in my WinISD model from 19cu ft to 9, keeping everything else the same, the freq response is altered quite drastically. The size is definitely needed and the driver responds to the extra volume. I'm sure they will indeed work well with a PR, but they also model nicely in a typical LLT enclosure with low tuning, and do benefit from the added enclosure size, unlike the 5200. Perhaps the 5200 was the driver you were thinking of?

[WillyD]

So, respectfully, I believe you're incorrect about the way WinISD handles Xmax.

No, you're incorrect. You do not use the peak-to-peak xmax value when modeling. The one-way linear xmax is the standard way of measuring "xmax" and that is exactly what it means.

Put 38mm in the xmax data field, because that is the max linear excursion one-way, and xmax is almost always given for one-way, not peak-to-peak.

Certainly what you wrote is true of TC-Sounds' TC-5200 line of drivers. The 18 incher in that line does not model well for an LLT at all, requiring a much smaller enclosure IIRC, not nearly enough to put in a big port. That driver would have to be tuned in a PR enclosure if one chose not to go sealed with them. Not true of the LMS-5400 18", though. For example, if I drop the enclosure size in my WinISD model from 19cu ft to 9, keeping everything else the same, the freq response is altered quite drastically. The size is definitely needed and the driver responds to the extra volume. I'm sure they will indeed work well with a PR, but they also model nicely in a typical LLT enclosure with low tuning, and do benefit from the added enclosure size, unlike the 5200. Perhaps the 5200 was the driver you were thinking of?

No, I was talking about the 5400. You're stretching what I said just a bit though. The 5400 does model well in a larger and lower tuned sub, but it doesn't need a 19ft enclosure.

The way the 5400 models, to port it properly in such a large/low tuned design you end up with way too low of a 1st port resonance, IMO. It is simply better suited to slightly smaller designs IMO, like 13ft^3 at 13Hz, or 14ft^3 at 12Hz.

[davepete]

WillyD wrote:

No, you're incorrect. You do not use the peak-to-peak xmax value when modeling. The one-way linear xmax is the standard way of measuring "xmax" and that is exactly what it means.

Put 38mm in the xmax data field, because that is the max linear excursion one-way, and xmax is almost always given for one-way, not peak-to-peak.

Xmax is often given as a one-way value, but that's an informal standard at best. There really is no standard, and one certainly can't just assume a particular program takes it one way vs the other. That's why it helps to do some manual calculations to see which way the program takes it. WinISD takes it as peak to peak. It is not doubling the Xmax to calculate the Vd, so you have to enter it as p-to-p. If you don't, you won't have the right value for Vd in your model.

No, I was talking about the 5400. You're stretching what I said just a bit though. The 5400 does model well in a larger and lower tuned sub, but it doesn't need a 19ft enclosure.

The way the 5400 models, to port it properly in such a large/low tuned design you end up with way too low of a 1st port resonance, IMO. It is simply better suited to slightly smaller designs IMO, like 13ft^3 at 13Hz, or 14ft^3 at 12Hz.

Yes, at 19 cu ft with a single 8" port tuned to 12hz, the 1st port resonance is a bit low at about 156hz, but not horrible. Steve's writeup recommends around 190hz, so it's a bit under that, but I don't think it's likely to be a deal-buster. If that's not good enough, by boosting the volume up to 22 cu ft and keeping the same tuning and port size, the 1st port resonance rises to over 184hz while only inducing about a third of a dB of resurgence in the shelf response around 13.5hz. This wouldn't be an audible effect and it gets the 1st port resonance right near the sweet spot.

[WillyD]

I am sorry, but are you saying xmax is an informal standard but Vd is set in stone as peak to peak displacement?

Thats a bit odd, don't you think?

Xmax is almost always given as the one-way linear excursion. I can't think of many examples where it isn't. Its not our fault that in WinISD, the formula used to calculate Vd is simply xmax x Sd, when usually it is 2 * xmax * Sd. I don't know why you think you need to put in the peak-to-peak xmax simply because of this.

"Having the right Vd value for your model" is completely irrelevant. That value it gives you for Vd is not used in the actual simulation. You only need to pay attention to what the simulation gives you (i.e SPL)


And you're correct about the resonance, 156Hz shouldn't be much of an issue. Do you actually have any intentions of building an 5400 LLT? That would be interesting..to say the least.

[SteveCallas]

Steve, I'm still trying to understand your "amp-limited" concept as you conceive it. How are you determining amp size for a particular application to make it properly amp limited?

As Willd touched upon, I don't want to feed it so much power that it can exceed the driver's xmax capabilities above tuning.....in the tuning range of LLTs, max excursion use should peak ~20hz. Now, if you ever drive the amp to the point of clipping, you know two things. First, turn it down a bit, and second, if you see clipping occur on a more than rare occassion, your listening habbits demand more subwoofage.

Hmm, it clearly states "peak" right on the input screen. In addition, it appears to use that number as though it was p-to-p when calculating the VD

Vd should be peak to peak, but for the sake of modeling something like an amp limited design, one is only interested in the one way safe travel limits.

I do happen to think that the LMS-5400 and possibly the upcoming 4100 are better suited to passive radiator subs because they seem to like smaller enclosures which would problematic to port properly. They don't need 19ft^3 to work well, you know? And in a smaller enclosure, they can handle more power (like a full 3200W).

You'd be giving up low end sensitivity for high end sensitivity and running more power through the coil, in which case you may as well just go sealed with a LT. The goal should be as much output as possible with as little power as needed in my opinion, but to each his own. The 4000 practically demands a large enclosure - I don't know what the deal with the 4100 is, but I thought it was just a 4000 with a more efficient motor due to a different magnet

As for port issues, were I using a LMS 5400 18", I'd probably go 480 effective liters with a 7" diameter port that is 32" long and flared and feed it with 1800 watts. Good luck trying to find a 7" diameter cylinder, but if you can, air velocity is fine as long as you aren't constantly pushing the full 1800 watts with low teen material, conservatively getting ~117db levels. If you are, you shouldn't be trying to do it in a single driver sub anyway.

I wouldn't recommend making the first resonance much lower than 190hz....at some point it's gonna become audible. Assuming an 80hz, 4th order crossover, having it much lower than 190hz might be getting too close to the edge. Remember, the LLT is supposed to be as much of a no performance compromise design as possible - that means in all regards. If you wanna judge a guy based on what he does with his own stuff, my first port resoannce models to be ~270hz, so.....

[WillyD]

I don't know what the deal with the 4100 is, but I thought it was just a 4000 with a more efficient motor due to a different magnet

No, thats only where the differences begin. Not only will it be far more efficient (close to or better than the TC-2000) but have more Bl, much more excursion (38mm 100% linear compared to the 4000s 30mm) with a mechanical excursion of close to/around 50mm one-way.

It'll have a tall-roll surround, a titanium cone instead of an aluminum cone, higher power handling...basically, it is far better than the 4000.

Also, it'll have a much lower Q, therefore it won't need ridiculously large boxes like the 4000 does. I suspect it won't need more than 10ft^3 for a good LLT design, but we'll see.

[SteveCallas]

Ok. More Bl and a lower Qts aren't exactly good things for a LLT, but time will tell.

[WillyD]

SteveCallas wrote:

Ok. More Bl and a lower Qts aren't exactly good things for a LLT, but time will tell.

Why wouldn't more Bl be a good thing?

And by lower Qts, I mean much lower than the 4000's, which is close to .5. The Qts should be around .3, I think.

[SteveCallas]

A high Bl and a low Qts typically makes it hard to have a smooth tie of port ouput with driver rolloff. Such cases result in a resurgence of output near tuning, or an underdamped port region. The RLp15 is ALMOST guilty of this. Here's a link talking about Bl - ported alignments get mentioned in the 9th paragraph. Such qualities - high Bl and low Qts - typically work better in smaller enclosures, which makes it more suited for sealed. But who knows how much the parameters will be shifted.

[WillyD]

I see.


Well we know they will be "shifted" from the 4000 greatly, if thats what you are referring too.

[Exocer]

I believe Willy was mentioning the LMS 4100 to have a significant improvement in Bl in comparison to the LMS 4000 which has an advertised Bl of just 11T*m. I don't think its fair to assume it'll suffer from issues with port/driver output roll off because it'll have a "higher" bl than that of the LMS 4000. For all we know, the final product may have a lower/similar bl to an Rl-p 15 which has already been proven to give pretty flat in-room response regardless of what simulations have shown about port output and driver roll-off.

Its fair to say the Rl-p 15 IS guilty of the phenomenon you mention Steve.The LMS-4100, if it measures closer to a TC-2000 in the Qts/Bl/ Vas will probably not suffer from this Fr linearity issues to a point where anyone would have to worry about it, especially since the Rl-p 15 LLT in-room measurements I've seen do not look like the simulations.

It would be very interesting to see a test rounding up a few drivers with certain power handling capabilities and see how the effect of VC heating effects the THD among these drivers. An educated guess would be to assume that Driver A with its higher power handling capabilities will probably be more resistant to VC heating side effects in comparison to Driver B with a lower thermal power handling capability.

[SteveCallas]

For all we know, the final product may have a lower/similar bl to an Rl-p 15 which has already been proven to give pretty flat in-room response regardless of what simulations have shown about port output and driver roll-off.

Its fair to say the Rl-p 15 IS guilty of the phenomenon you mention Steve

The RLp15 LLT design doesn't suffer from this because I accounted for it in the design. 260 effective liters with a ~15hz tune minimizes the discrepancy in the tie in of port output to driver rolloff enough that it isn't a significant issue.

especially since the Rl-p 15 LLT in-room measurements I've seen do not look like the simulations

That's because an it's an in room measurement room gain and/or reflections easily mask any small FR changes. To match the simulations, you'd want to do a ground plane measurement.....since that's not very easy, close mic measurements of the drivers and port with no crossovers in play is the next best option. Ryan happend to do something close to that, except he left the 100hz crossover in play.



You take out the crossover and the upper frequency bass will be even higher in magnitude. In the case of Ryan's specific design, max output should be somewhere in the neighborhood of 110hz. You combine the close mic driver response with the port output and you have a response that looks VERY similar to the simulation. All said and done, the simulations are very accurate. A ported sub response is no more difficult or no less accurate to predict than a sealed sub with an EQ boost (the EQ boost is replaced by port output).

Aside from the issues a resurgence in port output will create if left unattended to, the bigger problem is that such drivers that are prone to that want smaller enclosures to get away with lower tunes. This goes against being able to create an affective LLT, and this is why the RLp15 is ALMOST not a good choice. Much smaller than 260 effective liters and you'd fall into port troubles as well as reduce low end efficiency to the point that in room response probably wouldn't be as flat as it is.

[SteveCallas]

An educated guess would be to assume that Driver A with its higher power handling capabilities will probably be more resistant to VC heating side effects in comparison to Driver B with a lower thermal power handling capability.

Absolutely. The LMS drivers have HUGE coils with lots of turns of copper, so they can handle a lot of power and will be more resistant to THD created by vc heat than a driver with a small coil like one using xbl^2. That said, less power is still always going to be better than more power. Aside from the direct THD issues, one also has to wonder if high power handling in small sealed enclsoures could heat the internal enclosure air enough to expand the air, increasing the internal spring force, decreasing low end efficiency.

[davepete]

WillyD wrote:

I am sorry, but are you saying xmax is an informal standard but Vd is set in stone as peak to peak displacement?

Thats a bit odd, don't you think?

Yes, it is. I suppose Vd might not be set in stone, but I do believe it's typically calc'd using peak-to-peak excursion. In any case, I suppose we could check with LinearTeam to see which way they intended the Xmax to be entered. A definitive answer on this would be nice.

Xmax is almost always given as the one-way linear excursion. I can't think of many examples where it isn't. Its not our fault that in WinISD, the formula used to calculate Vd is simply xmax x Sd, when usually it is 2 * xmax * Sd. I don't know why you think you need to put in the peak-to-peak xmax simply because of this.

"Having the right Vd value for your model" is completely irrelevant. That value it gives you for Vd is not used in the actual simulation. You only need to pay attention to what the simulation gives you (i.e SPL)

Absolutely. I was only using Vd as a means to understand how to enter the Xmax.

And you're correct about the resonance, 156Hz shouldn't be much of an issue. Do you actually have any intentions of building an 5400 LLT? That would be interesting..to say the least.

Yes it would. I don't know that I'll do it using the LMS driver just yet, but I'll probably put an LLT together using either a 12" or 15" TC-2000 to evaluate it vs my sealed subs. I'm leaning toward using Sonotube in this effort, even though I've never built a sonosub before. The larger enclosures required for LLTs would make MDF enclosures extremely heavy and difficult to work with. So, for those of you who have built sonotube enclosures and don't mind offering up some of your wisdom on what works and what doesn't, I'd be interested in hearing it. Stuff like how to make an endcap. A Jasper Jig will cut a disk about 18" in diameter, but not 24" or 30". How is it normally done? I suppose I could rig something to use as a pivot for a router...