Does an (S)LLT require a subsonic filter?

[brucek]

I was wondering if the LLT or SLLT subs that people have been constructing here require a subsonic filter?

My thinking came from a feeling that a lot of electronics have very good low end response and can't be counted on as a natural rolloff at the extreme low frequencies < 10Hz......

I guess the prevailing sentiment is never.

The following posts are moved from the IB Makover thread. We were getting off track there.



brucek

--------------------------------------------------------------------------------------------

This project is looking like a huge success, you not only are getting more headroom, lower distortion, a flatter and deeper extending natural FR, and more impact, you are also minimizing the bass spillage into the rest of your home. Hopefully some will take the concept more seriously now

Basically, this is a large ported sub built into the wall, similar to Sonnies Behemoth, except its fixed by the fact that the wall is where the drivers are mounted, so it can't be moved.

Are you saying that conventional IB just isn't the way to go, because from the results here, it sure looks like it?

brucek

[Sonnie]

Are you saying that conventional IB just isn't the way to go, because from the results here, it sure looks like it?

At this point, that would sure be my conclusion. You could probably do just as well with a pair of drivers vs. four and more than make up for the extra expense of the wood. Of course Rodny is going for the super kill and I don't blame him, I would to if I had his setup.

[SteveCallas]

Yeah, I recently measured my reciever's sub out to be dead flat to 3.5hz - I'm going to have to revise that part of my LLT expained thread. So far though, I haven't heard any reports of bottoming from LLT owners, and this includes the ones straddling the upper limits with tunings in the 15-16hz range. In the designing stages of my sub, I thought a highpass was essential. I was going to buy an HK receiver because their EQ systems include a selectable highpass based on sub size. When I noticed my sub didn't bottom with really low bass scenes, I figured it was because it used an 18" driver, had plenty of headroom, and I didn't listen at reference levels in my apartment. As I started helping others design sus, some would opt to go for no highpass. Chuck (chasw) really stepped things up by using a ~17hz tune and 1400 watts with no highpass, a double whammie so to speak Yet even at levels as loud as he could handle, he reported no bottoming (after about a year of use he did report it was able to bottom when going to extremes). So after observing this with a few subs, I made the requirements of the LLT design such that I felt there was more than enough safety in play. It's kind of like a car manufacturer setting the hash mark for an empty gas tank when there are still a couple gallons of gas left.

[Ilkka]

brucek wrote:

I know you said the electronics will handle the rolloff, but I wouldn't count on it. The BFD has a flat response down to 5Hz and is only off a bit at 2Hz.......

That's strange, my measurements show much worse performance. Check below.

SteveCallas wrote:

Yeah, I recently measured my reciever's sub out to be dead flat to 3.5hz - I'm going to have to revise that part of my LLT expained thread.

If your receiver is truly that flat - congratulations (could you show us the measurement?). But don't assume that every AVR is alike. Here's a recent measurement showing the roll-off profile for my NAD T743 receiver (digital in - sub out) + Behringer DCX2496 (alone the DCX is similar to BFD), for the BDF, and for one pro amp that I have.

So for example my own system is 12 dB down at 5 Hz. That's similar to a 1st order L-R high pass at 10 Hz.

[brucek]

(could you show us the measurement?).

I have two BFD's. One 1100P and one 1124P.

I took a REW measure of two cables connected with an in-line short from 2Hz to 23KHz.

Then I removed the in-line short and inserted my 1100P with the signal passing through one channel (with the filters off). This eliminates all variables except the BFD response.
BFD set to -10dBV operating level.
Input level -10dB LED VU lit.

I'd say the response is pretty good. Only down a dB at 2Hz....

Here's two graphs.

The first graph is with a vertical scale of 45dB to 105dB as a reference showing the cable measure and the BFD response.

The second graph is the same with a much zoomed vertical scale to see small differences.

bfd response and cable loopback.jpg
ZOOMED IN.........
bfd response and cable loopback expand.jpg

So all that's standing between near DC response would be the processor. I'll bet my $6000 processor goes fairly low, but I've never measured it. If I had one of these LLT's, I would have be very careful. (I don't, I have a Servo-15.)

I suspect it's probably the source equipment and media that usually provides the protection. but, I think you have to be careful without a subsonic filter if you create one of these monsters.

brucek

[Ilkka]

brucek wrote:

I have two BFD's. One 1100P and one 1124P.

Could you measure the 1124P too? It would be nice to see whether it matches with my measurement.

So all that's standing between near DC response would be the processor. I'll bet my $6000 processor goes fairly low, but I've never measured it. If I had one of these LLT's, I would have be very careful. (I don't, I have a Servo-15.)

I suspect it's probably the source equipment and media that usually provides the protection. but, I think you have to be careful without a subsonic filter if you create one of these monsters.

brucek

Don't forget that you would a pro amp for the LLT. That would be another variable in the chain.

I'm not into betting, so if you can measure your processor, please do.

But in any case, different units definitely have different roll-off profiles. Some have more, some have less, best to measure them if you want to know.

[brucek]

Could you measure the 1124P too? It would be nice to see whether it matches with my measurement.
.....................................
I'm not into betting, so if you can measure your processor, please do

You're making work for me.

Anyway, I measured my 1124P and indeed it's not as good as my 1100P, but it still has quite a good response.

I labelled the actual measurements on a zoomed graph to show how far down the 1124P was at the low end.

I overlayed the 1100P on the graph for visual comparison to show that it was somewhat better.

CABLE vs 1100P vs 1124P RESPONSE
cable vs 1100p vs 1124p.jpg


------------

With respect to processors, there can be various results depending on the internals that are used in the measurement.

With my processor (Bryston SP2) there are three possiblities.

1) Analog input - signal must pass through analog to digital convertor (ADC), DSP, then digital to analog convertor (DAC). This would be considered worst case response.

2) Digital input - signal must pass through DSP, then digital to analog convertor.

3) Analog input - signal passes through internal analog preamplifer. This would be considered best case response.

I only tested condition 1) and 3) to show worst and best.

I tested from 2Hz to 200Hz for these tests, not interested above that.

I tested both the SP2 alone (bypass and ADC/DAC mode) and showed those graphs for interest and then continued the looped signal through the BFD1124P to show the overall response with the SP2 + BFD 1124P together in both SP2 BYPASS and ADC/DAC mode.....................

The BFD had the IN/OUT GREEN LED Filters engaged on with all 12 filters in OF mode. I also tested with the IN/OUT GREEN LED OFF and the responses were identical, so I only show with the IN/OUT LED ON and filters OF. The signal passes completely through the BFD circuits, but with no filters entered.

The SP2 certainly tracked the target 60Hz crossover that REW layed down.

BRYSTON SP2 SUB OUT RESPONSE 60Hz crossover using internal ADC & DAC. (75dB target shown)
cable vs sp2 SUB OUT adc-dac.jpg

BRYSTON SP2 SUB OUT RESPONSE IN BYPASS MODE. No digitization.
cable vs sp2 SUB OUT in bypass.jpg

BRYSTON SP2 + BFD1124P together (both Bypass and ADC/DAC shown)
cable vs sp2 + bfd1124p.jpg

I guess my conclusion would be that if 5Hz was below the tuning of an LLT without a subsonic filter, I wouldn't want to feed it too much material because I'm only down a bit better than -3dB at 5Hz...... agree, disagree?

brucek

[Ilkka]

brucek wrote:

You're making work for me.

Anyway, I measured my 1124P and indeed it's not as good as my 1100P, but it still has quite a good response.

I guess my conclusion would be that if 5Hz was below the tuning of an LLT without a subsonic filter, I wouldn't want to feed it too much material because I'm only down a bit better than -3dB at 5Hz...... agree, disagree?

brucek

Thank you! It's interesting that the 1124P is so much worse. I wonder how's the FBQ2496 then...?

While your processor doesn't have much roll-off, combined with the BFD it already provides quite considerable protection against over-excursion.

Here's a simulation of two subwoofers. Both are tuned to 15 Hz, the gray one has dead flat electronics, the yellow one has electronics similar to yours. You can see that with the same input power, the yellow excursion curve barely exceeds the Xmax (but stays well below ~40mm Xmech) while the gray one bottoms mechanically. One can safely input ~2x more power even with as little roll-off as your system has.

[brucek]

the yellow excursion curve barely exceeds the Xmax (but stays well below ~40mm Xmech) while the gray one bottoms mechanically

Huh, I'm surprised, but there it is in black and white yellow.

Steve will likely be interested in this stuff......

brucek

[Hakka]

Steve, what AVR do you use? My Denon 3803 rolls off at around 16hz.

[SteveCallas]

I use a Yamaha HTR 5890 and I knew I should have just saved a jpeg of the measured FR - why I didn't I don't know I'll probably get around to doing it again tommorrow.

brucek wrote:

Steve will likely be interested in this stuff......

It's looking like the FR linearity of electronics will vary from brand to brand and model to model. In the grand scheme of things, it has yet to be an issue though - I think we'll be fine.


On to more pressing matters - I believe Rodny put the S in front of LLT for "super"....I wonder if it should be for "structural"?

[Rodny Alvarez]

Here is a graph of the two IB purple, SLLT green to 5Hz no EQ...........................


IBVSSLLT.jpg

[brucek]

to 5Hz

What are you using for a meter/microphone calibration file below 10Hz?

IB purple, SLLT green ...........................

Yeah, your graphs have made the point I was trying to get at Rodny. The new SLLT ported design is acting just like a ported versus sealed would act.

Below the tuning frequency, the port has quickly unloaded and the driver cones are free to move as if there was no enclosure at all. At that point, you can drive those cones to their limit fairly easily. Any normal ported design would have a subsonic filter protecting it, but because these LLT's are tuned so low, the theory (as I understand it), is that the electronics will roll off the power to the driver by the nature of their own poor low end response.

Steve and Ilkka seem assured it won't happen. Since I know almost zero about speaker design, I would defer to their thoughts on this.

brucek

[willy-be]

The choice of an infrasonic Fb, in and of itself affords no special protection to a vented alignment. If the electronics pass frequencies below Fb to the driver it will unload. So it's a bit naive to claim these designs need no supplementary protection.

Threads where the rotary subwoofer is discussed, illustrate the amount of very low frequency content studios currently encode in DVD soundtracks. It's logical to assume trend will continue if not expand with the potential of the newest formats used on HD/BR DVD soundracks.

So it's probably unwise to make blanket statements to the effect that all infrasonic Fb vented designs are inherently protected by a combination of Fb and electronic roll-off.

Aloha,
Willy

[SteveCallas]

brucek wrote:

Any normal ported design would have a subsonic filter protecting it, but because these LLT's are tuned so low, the theory (as I understand it), is that the electronics will roll off the power to the driver by the nature of their own poor low end response.

That is one form. Available headroom and amp limiting is another. If we were using something like an average commercial 12" ported sub that happened to be tuned as low as a LLT, max headroom would be considerably less than the high excursion 15 or 18" drivers most of us use. It would be far easier to completely tax and bring to bottoming such a commercial sub without a highpass. And keep in mind that in Ilkka's model, that's probably at full power. If you model a 15" driver LLT tuned to 15hz and include something like a 1st order 10hz highpass to account for electronics rolloff, try lowering the power to obtain more common levels, as one's demands on a sub aren't always gonna be ~113db or so With the common RL-p15 LLT design for instance, if we take a look at 400 watts, which equates to an average ~109db levels, max excursion is never reached, let alone xmech. In Rodny's case, with 4 drivers, I wouldn't waste 1 second worrying.

Additionally, if an amp is rated for ~1000 watts and the design is made to keep the driver pretty much within xmech with 1000 watts, amp limiting comes into play. It's not that the amp can't pump out any more than 1000 watts, but the clipping lights should be activated when it's taxed that much and you should be able to hear distress from the sub. If you notice that amp clipping is even a semi common occurrence, you need to build another sub to both keep you better protected and eliminate compression.

You bring all these things together and you have protection to the point where bottoming just hasn't been an issue.

willy-be wrote:

The choice of an infrasonic Fb, in and of itself affords no special protection to a vented alignment.

Sure it does. If we have two subs in roughly the same size enclosure with the same power, one with a 14hz tune and one with a 25hz tune, and neither with a highpass, which is more likely to bottom with low bass movies? Powerful 10-20hz bass is much more common than powerful single digits bass.

willy-be wrote:

If the electronics pass frequencies below Fb to the driver it will unload. So it's a bit naive to claim these designs need no supplementary protection.

Again, we get back to headroom. Was Rodny's ported sub bottoming at 7hz at 80db? With 4 ported drivers, he can achieve ~122db in room from about 10hz on up without ever exceeding xmech at any frequency....probably even louder than that.

[Rodny Alvarez]

I played the scene "Irene" from Black Hawk Down, volume at -10db, at this volume it's pretty loud to me and this is with one box only! with the IB 103db and the SLLT 106db, the subs did not bottom out at all!( NO EQ) I read somewhere that on this scene it goes down to 7hz.
I'm using my ECM8000 and the UB 1002 and the REQW spl meter, I know that my ECM is not calibrated to 5Hz, maybe when Sonnie gets his mic back I can get a real measurement!!

[Ilkka]

willy-be wrote:

The choice of an infrasonic Fb, in and of itself affords no special protection to a vented alignment. If the electronics pass frequencies below Fb to the driver it will unload. So it's a bit naive to claim these designs need no supplementary protection.

Threads where the rotary subwoofer is discussed, illustrate the amount of very low frequency content studios currently encode in DVD soundtracks. It's logical to assume trend will continue if not expand with the potential of the newest formats used on HD/BR DVD soundracks.

So it's probably unwise to make blanket statements to the effect that all infrasonic Fb vented designs are inherently protected by a combination of Fb and electronic roll-off.

Aloha,
Willy

Steve already covered it pretty well, but I just want to say that the common rule that all ported subs MUST HAVE a high pass below the tuning frequency dates back to the days when 10mm was a huge Xmax value and anyone using more than a single 12" driver was considered crazy.

It should be noted that it isn't that easy to mechanically bottom a system having four or so long-throw 15"-18" drivers, especially with a program signal. Below the tuning frequency, those drivers behave like an IB subwoofer. And it's not like every IB would bottom when anything lower than 20 Hz is being present. It takes a lot more power to mechanically bottom those drivers than the simulation program shows. That's because the program assumes