A different look at horns vs direct radiators???

[DrWho]

Lower sensitivity does not necessarily correlate to lower
efficiency. When you pad things down, you're getting less energy out
because you're putting less energy in - which means the driver
isn't working as hard.

It also seems to me that the harder dB's to
achieve are at the extremes of the horn's bandwidth, not the midband.

My question then is this--> is the single driver, low-pass, high efficiency unit a BETTER engineering compromise than the two (or multi) driver, all-pass(?) unit.

Considering that the distortion will be lower and that the horn will have more control over the polars, then yes, the horn is a better engineering decision.

Btw, have you had the chance to hear the Jub LF or the KPT-XII-MB?

[John Warren]

Roy-

My original question restated-

Can I get the SAME or BETTER performance in a vented box using two 15"? Watts are cheap.

If so, then why bother with a Klipschorn?

jw

edit-I'd like to throw it out there as food for thought. Some of the new drivers manufactured today are quite good.

[DrWho]

Actually, the mouth of a horn may be beyond the physical dimensions of the horn's enclosure. Designers of horns apply what some call end correction to account for the true mouth size of foreshortened horns. The particle velocity of a sound wave at the physical mouth structure is quite high, and in the case of the Khorn this wavefront curves around the front face of the enclosure creating a virtual horn segment large enough to cause horn loading to below 40 Hz.

Do you have any data to support that?

Based on the modelling software and the few horns I've built, it seems that the low frequency cutoff for a horn is mostly dependant on the length of the horn. As you make the mouth of the horn smaller, you just get closer and closer to driving a straight tube. The first peak of the modal resonance with the tube is going to correspond with the same Fc of a proper horn. All that to say, how is making the mouth larger going to lower the Fc?

[Edgar]

Can I get the SAME or BETTER performance in a vented box using two 15"?

Roy can answer for himself, but I'll take the liberty of offering my opinion. There is more to performance than frequency response. For effortless, natural, undistorted, dynamic bass performance, I've never heard a DR that could approach a horn. Vented boxes are especially poor at this, in my experience.

Standard disclaimers apply: IMHO, YMMV, etc., etc.

Greg

[Don Richard]

Based on the modelling software and the few horns I've built, it seems that the low frequency cutoff for a horn is mostly dependant on the length of the horn. As you make the mouth of the horn smaller, you just get closer and closer to driving a straight tube. The first peak of the modal resonance with the tube is going to correspond with the same Fc of a proper horn. All that to say, how is making the mouth larger going to lower the Fc?

The LF cutoff of a horn is dependant on the flare rate. What you are referring to applies to open-ended and closed-ended resonators such as organ pipes. Horns used as loudspeakers are non-resonant in their designed operating range.

The Fc is not lowered. Making the mouth of a foreshortened horn the proper size allows that horn to load the driver better at lower frequencies.

http://ldsg.snippets.org/HORNS/design.html

This is one explanation of end correction as applies to horn loudspeakers. Googling end correction yields many hits, some having application regarding the organ pipes that you referred to earlier.

Edit: The concept of end correction in the design of acoustic horns is not a novel one. The treatises on acoustic horn design by Rayleigh and Olsen both mention it and provide formulae for the calculation thereof.

[jwc]

Btw, have you had the chance to hear the Jub LF or the KPT-XII-MB?

Off topic: I would like to hear from anyone who has heard the KPT-XII-MB? Dennis??

jc

[John Warren]

[

Roy can answer for himself, but I'll take the liberty of offering my opinion.

C'mon, what do you think he'll say?

I'm just hassling him.

jw

[DrWho]

The LF cutoff of a horn is dependant on the flare rate. What you are referring to applies to open-ended and closed-ended resonators such as organ pipes. Horns used as loudspeakers are non-resonant in their designed operating range.

So why does the ~20Hz flare of the horn I modelled drop off at 40Hz? If you crank through the simulation, you'll notice that any arbitrary flare rate will drop the frequency off at 40Hz when the length of the horn is ~2m (working off memory here) - provided the flare rate is below 40Hz of course. Well actually, the 40Hz flare rate drops off at like 50 or 60Hz (can't remember).

Also, every horn with an undersized mouth is going to resonate - it is a direct result of the impedance mismatch. I brought up the behavior of a pipe, not because that's how we want it to behave, but because that is the common sense check to see where the behavior will trend towards as the mouth is made smaller...or rather, the direction the trend moves away from as the mouth gets larger. The horn models assume that the mouth is extremely large in order to hold valid - so anything less than full size (which is what we're talking about here) cannot exceed that performance. In fact, anything smaller than "huge" is going to trend towards a resonate pipe.

Btw, this is the only snippet on end correction that I saw mentioned in your link:

Another factor to be aware of is that
the acoustic length of the horn may be substantially different than the
physical length due to the end correction. This depends on the shape of
the horn mouth, but it can add around 0.6 times the diameter of the
mouth to the acoustic length of the horn. In a short horn with a large
mouth, this effect could be substantial.

Not to be annoying, but that's not data...that's just some dude throwing out numbers. I'm not saying he's right or wrong, but that's not proof of concept.

How would one measure end correction anyway? Compare measured results against a prediction? Why not just call the prediction wrong in the first place?

[DrWho]

[

Roy can answer for himself, but I'll take the liberty of offering my opinion.

C'mon, what do you think he'll say?

I'm just hassling him.

jw

lol, that's essentially the very system I'm looking to improve upon []

Now I have to build the horn...

[jwc]

Bentz. What is you size limitation in HxWxD?

Also....will they be free standing? or will they be close to a wall or corner?

jc

[Don Richard]

Btw, this is the only snippet on end correction that I saw mentioned in your link:

Another factor to be aware of is that the acoustic length of the horn may be substantially different than the physical length due to the end correction. This depends on the shape of the horn mouth, but it can add around 0.6 times the diameter of the mouth to the acoustic length of the horn. In a short horn with a large mouth, this effect could be substantial.

Not to be annoying, but that's not data...that's just some dude throwing out numbers. I'm not saying he's right or wrong, but that's not proof of concept.

How would one measure end correction anyway? Compare measured results against a prediction? Why not just call the prediction wrong in the first place

End correction is not a novel idea. Rayleigh and Olsen mention it and give the math in their texts. They explain how end correction is different for the different types of horn flares.They also describe how an acoustic horn may be designed. There you will see that the flare rate of an acoustic horn is the primary determinant of LF cutoff. This is confirmed in the more recent writings of Klipsch and Leach on that subject. Note that if you follow the different authors' methods of horn design you will get slightly different final results. This is because a horn may be optimized for varying design objectives such as efficiency, size, bandwidth, smoothness of response, maximum output, pattern control, etc. The early horns were designed for maximum efficiency to go with the low power amplifiers of the day. This generally results in large sizes and frequency response limitations.

Horns work best when certain ratios are observed. One rule of thumb that I have seen in more than one place is that a horn's path length exceed 1/4 wavelength of the lowest frequency to be reproduced. When shortened too much the device is no longer a horn and cannot be expected to have the characteristics of a horn. This may help explain the results you weren't expecting.

Regarding your original post, two of the best ways to achieve good results for your application would be a bandpass-horn enclosure or a conical bass horn. In a bandpass horn one balances the back chamber volume with a front chamber volume like a bandpass subwoofer. The output slot is then horn loaded. For the best writeup on conical bass horns:

http://www.danleysoundlabs.com/pdf/danley_tapped.pdf

Read the section on Synergy Horns for design info and pictures of what loading a driver near the mouth of a conical horn looks like.

Most concert horn loaded enclosures for bass duty these days use either bandpass horns, conicals, or tapped horns. Turbosound, EAW, Yorkville, Danley are some of the companies using these types of bass horns for sound reinforcement. They are not as efficient as the old exponential horns but they are still more efficient than direct radiator enclosures and they take up less room than vented enclosures for the same output. Hornresponse easily models bandpass horns, but the tapped horn and conical bass horn design features are lacking or incomplete.

[ClaudeJ1]

I dunno what I'm gonna do.

To be honest, I know what the vented cabs will do, but I'm worried about screwing up the horn and I'm not sure I can guarantee outperforming the vented cab (unless I go rediculously huge with it).

Why don't you just build a couple of Tom Danley's LAB horns? You could do that for less than $1,000/pair and they get to 30 Hz with 106 db efficiency in a 16 cubic foot enclosure.

I am trying to reuse the four JBL2226J drivers that I'm tearing out of the beat up vented cabs.

The Lab Horn is designed for the Eminence Lab12 (12" driver) and requires something like 4 cabs side by side to achieve usable output to 30Hz. Is that the one you're talking about?

YES. Here's Tom's post from the High Efficiency Speaker Asylum from Feb. 26, 2003.

I am glad you are pleased with the driver, I think you will be more so when you finish the horn part of the project.

The driver is a good one, it does work well and I am sure it will be
made for a while as I know Jerry at Eminence was building
the LAB horns too.
Part of the reason for my designing the LAB horn and specifying the driver parameters was to give the Pro-sound DIY'rs a
modern bass horn, so many people interested in horns still follow the old " folk lore ", like for example that a bass horn driver
should have low mass and small excursion capacity.

When a couple of people on the live audio board suggested using a JBL-K
series driver I stepped in and said look, if your going
to all the trouble to build something like a horn, why not make it work as well as it can and not start life crippled.

In reality, the proper driver parameters will always be the best choice
and not ones chosen by "thumb rule" so while this driver
seems very heavy to some, those who measure the box find a different story, it works exactly as intended.

At some point, the folks who did a shoot out in Michigan will have the
measurements and listening results posted on the basslist
so more actual data will be available soon.

Anyway, I am pleased that some HIFI types have picked up on the design,
Don B was one of the first I was aware of to use it in
a home. I
was inspired myself, I have always had Servodrive subs around usually
(often prototypes) so I am not often lacking in
bass extension, my in room listening position response is pretty flat to 10 Hz.
On the other hand, I have rarely had any of our big horns at home, they seemed "over kill" when a ContraBass or two can scare
strangers..

I had just designed a variation on the LAB sub horn and got it in my living room last weekend, a little smaller, only 42 by 42 by
18 inches which also uses a "trick" in its operation. In room, it is unusually smooth for a small horn + - 3 dB from 30 to 225 Hz
and with 10 dB of eq (no problem because of high sensitivity ~ 106 db 1w) the measured response is - 3 dB at 22 HZ.
Using the TEF machine to measure phase and the Omnidrive (BS366) to mate them, I got it aligned with the Unity horn system
on top.

Listening results:
Well, I am re-thinking my furniture arrangement in the living room to accommodate 2 LF horns (a right and left).
The "closer to zero" degree acoustic phase of the LF horn shows up sounding very "tight" compared to the direct radiator's ~-90
degree phase (which is in effect an increasing delay with decreasing F).
At the moment, the increased headroom and the fact that I can mate the acoustic phase with the Unity above (resulting in near
zero degree acoustic phase from about 50 Hz to 5 KHz) makes the horn woofer more desirable and eq buys back some of the
LF extension. At a loud listening level, the LF THD was 2% or less from 25 to 150 Hz so compared to a direct radiating cone
driver, the horn is also much lower in distortion.

No question, I think you will like those drivers in a horn.
Are you going to the horn fest in Lima?
Cheers,

Tom Danley

[jwc]

So why does the ~20Hz flare of the horn I modelled drop off at 40Hz? If you crank through the simulation, you'll notice that any arbitrary flare rate will drop the frequency off at 40Hz when the length of the horn is ~2m (working off memory here) - provided the flare rate is below 40Hz of course. Well actually, the 40Hz flare rate drops off at like 50 or 60Hz (can't remember).

Edgar and Bentz

Model this? Not sure if this will fit in your "space" Bentz.

Mouth (exit) 7580 sq cm

Horn length (throat to exit) 122.7cm

Throat 870 sq cm

Put it in a corner.

What is the T/S of those JBL's? What is the total diameter of those JBL's??? 15.25"??

jc

[boom3]

Hi Don Richard

Your statement:

"...and in the case of the Khorn this wavefront curves

around the front face of the enclosure creating a virtual horn segment

large enough to cause horn loading to below 40 Hz"

is not supported by anything I ever read by Paul or his proteges. In fact the two sides of the horn cause an intereference dip at some frequency. The front face of the Klipschorn is far too small to be an effective baffle at 40 Hz. If you are saying that the K-horn (plus corner) becomes the throat of a "virtual" 40 Hz horn formed by the room, then we must specify room size, wall stiffness, etc. In fact, even if the throat/driver could be tunnelled back into a room corner, a 90 degree room corner forms a conical flare, which would greatly reduce low-band efficiency.

If you read Paul's "Note On Acoustic Horns", (Proceedings of the IRE, Vol. 33, No. 7, July 1945) As usual, the maestro has given us some clues about performance of horns below cut-off.. "Last but perhaps not least, is the fact that the resistive component remains appreciable for a considerable range below cut-ff, so that some useful contribution of sound energy is avaialble from large-throat speakers". So, it seems that a woofer with a back chamber resonated above nominal horn cut-off is acting like an "infinite baffle" below the horns cut-off and that is why we still get usable bass out of horns below nominal cut-off.

[Marvel]

What is the T/S of those JBL's? What is the total diameter of those JBL's??? 15.25"??

It is the 2226J, listed on this page.

Bruce

[jwc]

Thanks Bruce.

The driver has a diameter of 15.25" too.

SO! you software nuts run the numbers above. I have sketched out the cuts for a horn for two JBL 2226 drivers. I was off a little on the back chamber volume. So......the horn will be a little longer and the mouth will be a little bigger.

It's gettin big though.

jc

[Edgar]

So why does the ~20Hz flare of the horn I modelled drop off at 40Hz? If you crank through the simulation, you'll notice that any arbitrary flare rate will drop the frequency off at 40Hz when the length of the horn is ~2m (working off memory here) - provided the flare rate is below 40Hz of course.

For what it's worth, this has been my experience, as well. Recently I've been designing a corner horn about 1.9 meters long. No matter what I do to the contour, the response cuts-off somewhere in the region of 45 Hz. The contour only affects the smoothness of the response.

Basically it seems that a horn ain't a horn until it's at least 1/4 wavelength long.

Greg

[sfogg]

"
Basically it seems that a horn ain't a horn until it's at least 1/4 wavelength long."

As an example to the contrary though the K510 is roughly 5.25" deep and is good to 500hz. 1/4 wavelength of 500hz is roughly 6.75".

I think Roy has a stash of Alice's 'Drink Me' juice though. Wonder what would happen if he used that on a 80hz straight horn?

Shawn

[boom3]

Greg wrote:

"Basically it seems that a horn ain't a horn until it's at least 1/4 wavelength long."

***

Right! The length of a horn can't be fudged. The size of the mouth, for given loading (corner, wall, etc) will vary but it has to be a minimum size. Mike's comment about end correction is not entirely without foundation. Don Keele has written a paper about that which I don't have at my fingertips but is on Keele's website. In sum, a mouth can be too big for a given size of horn. This is not a license to try to cheat physics, it is a caution against going overboard.

[Edgar]

As an example to the contrary though the K510 is roughly 5.25" deep and is good to 500hz. 1/4 wavelength of 500hz is roughly 6.75".

Roy knows how to warp time.

Seriously, though, don't forget that there's some horn length attributable to the depth of the compression driver.

Greg