A different look at horns vs direct radiators???

[DrWho]

So I've got this interesting opportunity to build some rather large
bass horns on campus, but have run into an interesting problem...I can't seem to get more performance by horn loading the system []

I
used to work for a venue on campus and my boss was hoping that I'd
rebuild some of the old worn out equipment before I was totally gone
for good. Long story short, I have four JBL 2226J drivers that I would
like to hornload. They were previously wired in parallel in 8 cubic
foot ported enclosures tuned to 40Hz. This essentially yields a system
sensitivity of 103dB (since each driver is normally 97dB) that is
essentially flat down to 40Hz.

I would like the hornloaded system to maintain the same bandwidth.

Crunching
through the models in Hornresp, it looks like I can get around an
additional 6dB increase in efficiency. However, I can't seem to get the
horn to take up less than 24 cubic feet. So looking at the size of the
cabinets, I should be able to yield more output per cubic foot from the
vented cabinets...

While acoustic output isn't my determining
factor of performance, I just don't see how the horn is going to be
lower distortion than a pair of vented cabs. The cone excursion
for the horn and the dual vented cabs is effectively the same for the
same SPL, so driver distortions should be extremely similar. And then
you have all the reactance in the horn that you don't have in the
vented cab...

Am I missing something obvious here?

I
suppose I only have four drivers, so a pair of horns are going to be
louder (or less distortion at the same SPL) than a pair of vented cabs,
but I think it just kinda sucks that they're going to be so big.

For
what it's worth, I was thinking of making them straight horns too (so
not folded), which I'm hoping will reduce the complexity and get me
closer to the Hornresp model.

Here's the hornresp output...dark
line is the horn and the light gray line is the ported cabinet. Both
systems are using dual 15" drivers.

[dgoreck]

Pardon my newbiness to hornloading and such, but when I built a few custom boxes for my car subwoofers, I remembered, that different drives would produce way different results in a vented vs sealed enclosure. I knw the Polk Audio db10's I had did not do well when vented, but where awesome in a sealed box. Thi has to do with the design of the subwoofer itself I thought. Would the same kind of principles apply to hornloading them, folded or not?

[DrWho]

I 'think' the driver selection only affects the rear chamber volume (provided of course the drivercan go low enough to begin with). In this case it is 120L for the 2 drivers.

[ClaudeJ1]

Keele agreed with your findings about 33 years ago.

KeeleOnHornsVSPortedDirectRadiator.pdf

KeeleOnHornsVSPortedDirectRadiator.pdf

[ClaudeJ1]

Woudn't it just be easier to build some MWM bins?

[Edgar]

Am I missing something obvious here?

Nope. See Don Keele's paper, especially the conclusion.

Greg

[jwc]

Mike. I am not ashamed of going the bass reflex route as I've messed with both including the DIY.

Noce to see SOMEONE post interest in bass horn DIY.

Can you help me by giving me more input of what your input is into HornResp.

Is this 1/2 space?

Is this pure exponential flare?

What was the horn length, total throat and mouth area of the "horn" with the curve shown above?

The only way I know to get your cabinet size down in size and achieve a curve like what you have....is to make it an 1/8 space horn. I wouldn't be to thrilled with that and would choose the bass reflex option. I also don't know how you manage to get such a "flat" response from 40-100Hz. Every exponential 40Hz horn I've seen curves on drops about 6-10 dB. Unless there is some hyperbolic component.

How about MWM cabinets calculated to 1/4 space.

jc

[DrWho]

Here's the horn parameters:

The throat is 880 sq cm, which is the surface area of one driver (I'm using two drivers though). The mouth is 6600sq cm. I'm using a 24Hz cutoff frequency on a tractrix area expansion, which comes to a horn length of 230cm.

You get about the same bandwidth with the tractrix as you do with the exponential and conical horns, but the tractrix seems to have less dramatic peaks and dips.

The rear cabinet volume is 120L for the two drivers (about 4 cubic feet). Changing the rear volume has a dramatic effect on that last peak at 40Hz and the dip at 50Hz right before it. I think this is what everyone is referring to when they talk about reactance annulling. As you make the rear chamber smaller, you can raise the output at 40Hz, but at the expense of less output at 50Hz.

I'm modelling everything in 1/2 space because that would be the worst case scenario is a pro-sound environment. I always double check what 1/4 and 1/8 space look like though, because sometimes you can make something look good in one, but will be crap in the others. This horn actually models ruler flat to 40Hz in 1/8 space.

Here's the Hornresp input params:

[DrWho]

Btw, I am familiar with Keele's paper on this subject, but I'm not sure if I agree with it entirely...

Fundamentally, there is nothing special about specific frequency ranges in acoustics. In other words, something that is true for one range of frequencies is going to be absolutely true at another range of frequencies...the implication that Keele's article could be extrapolated to also apply to squawkers and tweeters. PWK and crew wouldn't be such large fans of horns if there wasn't more to it...

I've heard Roy mention the polar response just about everytime Keele's article is mentioned. I just wonder how significant it is when you're talking only an octave of bandwidth (especially when the wavelengths are so much larger than the drivers). I also don't have the means to reliably predict the polar response, and I can't exactly be building a few dozen prototypes to start figuring out the trends emperically...

The real question I have though, is if the four ported drivers are going to have less distortion than the single dual driver horn - the acoustic output is the same and the driver excursion is about the same...and I see no reason why the polars shouldn't be about the same either. To me, that means that if the horn isn't perfect, that it is going to be adding distortions in addition to the driver distortions - the net effect being that the four ported drivers will be better. I only have four drivers to play with right now, but I might be inclined to purchase four more if it's going to cost about the same as all the wood needed for the horn.

[jwc]

Mike. I follow you now.

Yes with a tractrix flare, a horn cutoff of 24 Hz doesn't necessarily lead to a gigantic horn. But. Do you have any literature regarding the use of tractrix with such low freqeuncies? Bruce Edgar touched on this concept some and favored against the idea but this seemed more of "Point" rather than something he experimented with.

I was going to build a big tractrix horn........but I got to wondering about the "folding" aspect of tractrix. Is there any more compromise of folding tractix vs exponential. Don't know. I didn't want to break my back on a big Tractrix horn with the unknowns. Heck, one/two fold of a tractrix bass horn shouldn't be that hard of a build.

jc

[PrestonTom]

Mike, Let's just speculate for the moment. In the Jubilee bass bin, the back volume requirement is about 30 or so liters per driver. The horn is foreshortened and the mouth is undersized (relative to the flare). Why not extend the cabinet and give it something like a La Scala nose and extend the outside also . You would probably still be within your required footprint. You could substitute the pioneer driver of similar T/S specs and they only cost about $50/driver.

Essentially give the jub bass bin the "proper sized" cabinet. Will that go deep enough for you? An adventurous soul might actually try adding the passive radiator to the design (as was originally "played with"). But that would require some honest-to-goodness engineering....

Regarding the Tractrix flares: I have drawn out some and that get very big very quickly (the mouth that is, the length actually gets a bit shorter). Although horn guys frequently foreshorten and under size exponential horns (so they can fit through a doorway), I have no idea if you can play the same game with tractrix horns. Hmmm ......

Simply food for thought,

-Tom

[DrWho]

Yes with a tractrix flare, a horn cutoff of 24 Hz doesn't necessarily lead to a gigantic horn. But. Do you have any literature regarding the use of tractrix with such low freqeuncies? Bruce Edgar touched on this concept some and favored against the idea but this seemed more of "Point" rather than something he experimented with.

All I know is Roy says Edgar needs to see some real data []

Do you happen to have that article where Edgar says tractrix isn't good for bass horns? I wonder if his argument doesn't stem from the fact that most bass horns tend to be as undersized as possible...making the horn smaller relative to the wavelengths

The tractrix area expansion equation assumes that the wave starts off as a plane-wave in the throat and then morphs into a spherical wave at the mouth. I wonder if Edgar isn't arguing that most bass horns are too small to support a full spherical wave at the mouth. In which case, the wave is still going to look like a plane-wave at the exit. I also wonder if this isn't the "modified-tractrix" that Roy is always talking about...I wonder if Roy is assuming that it goes from planar to somewhat spherical, but not completely?

The way I see it - the arguments being made by Edgar and Keele and everyone seem to always be coming from the same math that the models are based on. So if the model shows a certain behavior, then there really isn't anything to be learned from their conclusions. I say that hesitantly because I don't want to imply that we shouldn't read their conclusions, but more often than not, most conclusions tend to be broad and general, which don't necessarily apply to every specific situation.

In other words, I see no reason not to believe the model based on conclusions drawn from the same math that these guys are using.

I was going to build a big tractrix horn........but I got to wondering about the "folding" aspect of tractrix. Is there any more compromise of folding tractix vs exponential. Don't know. I didn't want to break my back on a big Tractrix horn with the unknowns. Heck, one/two fold of a tractrix bass horn shouldn't be that hard of a build.

Why would there be more of a compromise between folding a tractrix and exponential? They're just area expansion rates???

[DrWho]

Mike, Let's just speculate for the moment. In the Jubilee bass bin, the back volume requirement is about 30 or so liters per driver. The horn is foreshortened and the mouth is undersized (relative to the flare). Why not extend the cabinet and give it something like a La Scala nose and extend the outside also . You would probably still be within your required footprint. You could substitute the pioneer driver of similar T/S specs and they only cost about $50/driver.

Essentially give the jub bass bin the "proper sized" cabinet. Will that go deep enough for you? An adventurous soul might actually try adding the passive radiator to the design (as was originally "played with"). But that would require some honest-to-goodness engineering....

Regarding the Tractrix flares: I have drawn out some and that get very big very quickly (the mouth that is, the length actually gets a bit shorter). Although horn guys frequently foreshorten and under size exponential horns (so they can fit through a doorway), I have no idea if you can play the same game with tractrix horns. Hmmm ......

Simply food for thought,

-Tom

Are those Pioneer drivers really only $50? []

So has anyone ever tried to put the Jub into Hornresp? I think it'd be interesting to compare the model against all the measurements we have available. I'm sure the folding smears things around a bit, but we should be able to get somewhere in the ballpark.

Building a Jub clone with the full horn is a real tempting idea. Sadly, I don't think it would be able to fit through doors if it were all one piece though. If one unfolded it to obtain more feasible dimensions, then I think you might as well start the design from scratch (of course applying everything learned from the Jub)?

Btw, doesn't nearly every Klipsch bass horn have an expansion rate lower than the actual Fc of the horn? I seem to recall that most Klipsch horns have undersized rear chamber volumes too? Roy says the Jub is a tractrix expansion and the horn has been cut too short too...

I'm thinking about playing around with multiple expansion rates next. If you look at the acoustical impedance in the model, it looks a lot like what you'd see on a pipe. I'm wondering if I can't spread out the Q of those impedance peaks, which I would imagine should lead to lower distortion and better loading in the nulls.

[Edgar]

All I know is Roy says Edgar needs to see some real data

Do you happen to have that article where Edgar says tractrix isn't good for bass horns?

Here.

II wonder if his argument doesn't stem from the fact that most bass horns tend to be as undersized as possible...making the horn smaller relative to the wavelengths

Actually, I think it's the opposite. With very short horns it is difficult to distinguish between exponential, tractrix, hyperbolic, or even conical. (Try your Hornresp model with a conical horn instead of a tractrix, and you'll see only small differences in response.) I think that it is only when the horns become long enough for the major differences in their contours to emerge that the response differences start to become significant.

Greg

(No, I'm not Bruce Edgar. Edgar is just my forum name. Maybe I'll change that someday, to avoid confusion.)

[Edgar]

So has anyone ever tried to put the Jub into Hornresp?

Yes.

[djk]

"the arguments being made by Edgar and Keele and everyone seem to always be coming from the same math that the models are based on."

The Keele article is not an 'argument' per se, it presents data from the EV TL5050 (dual 12 horn) vs the TL606Q (quad 15 reflex).

The reflex box is smaller, and eats the horn box for output below 70hz.

I have done side-by-sides of the classic 80" wide dual 15 theater horns (Altec A4/JBL4550/RCA Ubangi) vs the EV TL606D (dual 15 reflex), no contest below 100hz.

What is interesting is the EV TL606D vs the Altec A5/JBL4560 (single 15 horn).

On paper they have the same Di(Q), and about the same sensitivity.

The midrange sounds much better on the horn, the reflex is much better in the bass.

The compression drivers don't like the 500hz crossover, the 1:1 throat 15" cone driven horns don't like to go above 500hz.

The solution?

2:1 throat 10" or 12" cone driven horn mids with reflex bass. We can cross the mid horns high enough to select the HF drivers based on HF performance, not how well they work below 1Khz, and we can make the mid horn just big enough to do the job and still be easy to move.

JBL4560 on top of EV TL5050

Replacement PA (circa 1981)

Current modular mids with EV12, usually used in stacked pairs on top of Push-Pull dual 15s.

[Don Richard]

An option could be to get away from an exponential bass horn and employ some sort of combination enclosure.

http://www.speakerplans.com/index.php?id=hd15horn

This is a bandpass horn similar to the Turbosound TurboBass module. You are sacrificing upper end extension for bass power and smaller size, and these boxes need to be run in multiples (4 or 6) to get enough mouth area to be maximally effective at lower frequencies. These are for a 15" driver but Hornresponse can be used to design a 12" version. These are about the simplest horn enclosures to build that I have seen.

There is also a website that has plans for a LaScala type cabinet for a driver smaller than 15". I think that the website is Volvotreter.de, but not sure.

Good luck!

[Edgar]

Building a Jub clone with the full horn is a real tempting idea. Sadly, I don't think it would be able to fit through doors if it were all one piece though. If one unfolded it to obtain more feasible dimensions, then I think you might as well start the design from scratch (of course applying everything learned from the Jub)?

You can fit an almost perfect 39 Hz tractrix contour inside a Jubilee. I don't know that I'd want to do all that woodcutting, though. Hornresp says that it only subtracts a few Hz from the low frequency cutoff.

Greg

Edit: revised to say "subtracts from" cutoff instead of "adds to".

[Marvel]

There is also a website that has plans for a LaScala type cabinet for a driver smaller than 15". I think that the website is Volvotreter.de, but not sure.

This has even less bass response than the full sized LS. Erik uses a KHorn style bass bin for his low end.

Bruce

[Don Richard]

also don't have the means to reliably predict the polar response, and I can't exactly be building a few dozen prototypes to start figuring out the trends emperically...

Hornresponse's SPL window shows the power response of the horn. It can also give axial SPL at any polar that you select. After you calculate, go to tools, select directivity, then select response. A window will pop up with a place where you can enter the degrees off axis you wish to measure. Note that at 0 deg. the curve given will be the axial SPL, which is what you normally see on a mic'ed response curve of a speaker.

The Hornresponse program is an amazing program with many uses and features.

http://www.diyaudio.com/wiki/index.php?page=Hornresp+Help

This may help in understanding all the features this wonderful program offers.