Square Pegs

As far as the Iwata horns, if one is not concerned with constant directivity, a relatively low cutoff is not hard to achieve.

The Paraline which I posted polars for in post 496 has usable response down to 167 Hz (EV DH1AMT driver), and is only about 11.5 inches (29 centimeters) in length. It holds a uniform in horizontal response of 90 degrees from 500 -16 kHz, the much longer Iwata 300 goes from about 100 degrees to 15 degrees over the same range.

Art

Hello Art,

The design shown in message #497 is not an Iwata horn it is a very different design! The Iwata while having a lower horizontal directivity compared to the axisymetrical Le Cléac'h horn is not designed to control directivity. The new horn in #497 was designed in order to follow the current trend for controled directivity (even if for myself and others that want to benefit of a real 3D imaging from good stereo records there is no need for controled directivity as we listen only at the sweet spot andin the near field...)

What is seen in the frequency response of the Paraline you gave is that the ripples in the response curves could result in (multipaths) reflexions/diffractions inside the device, something that is avoid in Le Cléac'h horns.

Best regards from Paris, France

Jean-Michel Le Cléac'h
 
Originally Posted by weltersys

The T-35 horizontal dispersion goes from 110 degrees at 3kHz to 85 at 10kHz, a 25 degree range, the horizontal from 160 to 60, a 100 degree range. That "seems" a rather large difference to me.

I was looking at the circles & triangles.

The Paraline which I posted polars for in post 496 has usable response down to 167 Hz (EV DH1AMT driver), and is only about 11.5 inches (29 centimeters) in length. It holds a uniform in horizontal response of 90 degrees from 500 -16 kHz,

Impressive specs indeed, for such a short length :) But a +- 7dB fr to 1kHz = ?
 
I was looking at the circles & triangles.
Impressive specs indeed, for such a short length :) But a +- 7dB fr to 1kHz = ?
I also was looking at the circles and triangles which is why I noted the T-35 horizontal dispersion has at 3kHz to 85 at 10kHz, a 25 degree range, while the vertical has a 100 degree range in the same pass band, a rather large difference.

As Jean-Michel Le Cléac'h said in post #501:
"What is seen in the frequency response of the Paraline you gave is that the ripples in the response curves could result in (multipaths) reflexions/diffractions inside the device, something that is avoid in Le Cléac'h horns."

The Paraline ripples and pattern flip are something that can be easily avoided using a conical horn with flared throat and mouth, (though a 90 degree conical won't go as low) which I have pointed out several times in this thread.

As far as the Paraline's low frequency response, one can see that the Iwata- 300 you linked in post 498 has around +/-16 dB response from 200 to 1000Hz, while the Paraline is only +/-6 dB in that range.

Whether the horn is half empty or full of it depends on one's point of view;).

Art
 

Attachments

  • Iwata:Paraline.png
    Iwata:Paraline.png
    285.6 KB · Views: 1,146
Hi, been a while since I've been able to do any speaker projects due to other obligations, but I've recently dug out an older idea I had and was hoping some of the people more knowledgeable than me might be able to help me out.

The idea I have in mind is a sort of modular 2-way unity style horn with paraline HF entry. Main design criteria is highest possible voltage sensitivity from a small box, so it can be run on battery power in it's most mobile configuration. For higher SPLs it can supply a bit more voltage from a power connector, and be stacked in multiples. That's the way I imagine it anyway. It's inspired in part by the VTC EL210 as well as a design I've seen on the forum many moons ago, which I'm pretty sure was Art Welter's, but I haven't been able to find via search.

My idea is to use two Eminence Alpha(lite) 6As in parallel for the midrange (low 100s through about 2kHz) and a fairly cheap compression driver above that. Sensitivity should be at least 100dB at 2.83V, though each individual box doesn't need a huge amount of power handling.

Here's a rough sketchup of what I am thinking
GoVfpg2.png


Outer dimensions are 20cm high x 60cm wide x 40cm deep (8x24x16 inches)

My thought was that the paraline + horn shape should make for a fairly wide horizontal dispersion pattern, with a narrow vertical dispersion. However in light of what I've read here I'm concerned about diffraction at the long edges and pattern flip from the quite small vertical dimensions when used as a single box instead of a line array. Unfortunately I'm very inexperienced when it comes to horn directivity issues, and my simulations in hornresp only go so far (Only allows a single segment horn, only considers midrange, only allows for symmetrical horns, etc) - I've tried using multiple simulations, one with a 90 degree expanding conical horn, and one with 0 expansion parallel tube, in order to judge the directivity response in both axes seperately, but again, I still have no idea what the tweeter is doing here.

Any thoughts on this build concept? Oh yeah, this is paraline related, probably going to use 4mm or 5mm thick material for that, to allow wave lengths up to 20k to pass through, assuming that doesnt constrict too much. But I'm a bit more concerned how the directivity of a final unit is going to wind up.
 
Hi, been a while since I've been able to do any speaker projects due to other obligations, but I've recently dug out an older idea I had and was hoping some of the people more knowledgeable than me might be able to help me out.

The idea I have in mind is a sort of modular 2-way unity style horn with paraline HF entry. Main design criteria is highest possible voltage sensitivity from a small box, so it can be run on battery power in it's most mobile configuration. For higher SPLs it can supply a bit more voltage from a power connector, and be stacked in multiples. That's the way I imagine it anyway. It's inspired in part by the VTC EL210 as well as a design I've seen on the forum many moons ago, which I'm pretty sure was Art Welter's, but I haven't been able to find via search.

My idea is to use two Eminence Alpha(lite) 6As in parallel for the midrange (low 100s through about 2kHz) and a fairly cheap compression driver above that. Sensitivity should be at least 100dB at 2.83V, though each individual box doesn't need a huge amount of power handling.

Here's a rough sketchup of what I am thinking
GoVfpg2.png


Outer dimensions are 20cm high x 60cm wide x 40cm deep (8x24x16 inches)

My thought was that the paraline + horn shape should make for a fairly wide horizontal dispersion pattern, with a narrow vertical dispersion. However in light of what I've read here I'm concerned about diffraction at the long edges and pattern flip from the quite small vertical dimensions when used as a single box instead of a line array. Unfortunately I'm very inexperienced when it comes to horn directivity issues, and my simulations in hornresp only go so far (Only allows a single segment horn, only considers midrange, only allows for symmetrical horns, etc) - I've tried using multiple simulations, one with a 90 degree expanding conical horn, and one with 0 expansion parallel tube, in order to judge the directivity response in both axes seperately, but again, I still have no idea what the tweeter is doing here.

Any thoughts on this build concept? Oh yeah, this is paraline related, probably going to use 4mm or 5mm thick material for that, to allow wave lengths up to 20k to pass through, assuming that doesnt constrict too much. But I'm a bit more concerned how the directivity of a final unit is going to wind up.

Some thoughts:

1) What's the application?
2) Personally, I am not as concerned about the dips and peaks in the frequency response as most people are. For instance, if you make a call on your phone, the frequency response is far from flat. But the sound is not offensive. Now try making the phone call on a VOIP service like Skype. Even the smallest 'blip', a fraction of a second, is audible and offensive. That's how our hearing mechanism works; it's very sensitive to timing.
3) Pattern flip makes horns sound 'wonky' it's true. But there's an easy solution to pattern flip with the VTC boxes; just array them like they're supposed to be used. In your application that may be an acceptable solution.
4) Instead of investing a lot of time with Akabak and Hornresp, I'd simply download the dimensions of the VTC boxes and start from there. They've published all of the data in CLF format, so you can see *exactly* what the frequency response and directivity is like.

Having said all that, the main reason I like Paralines is that it allows you to package the horn in some very odd shapes. For instance, under the dash of a car, or have the horn fire one direction while the driver points another. In your app, a plain ol' Synergy horn might give you better results. I don't see that a Paraline would offer any efficiency advantage, as a fair amount of output is lost on the high end due to imperfect combining of the high frequencies. You'll notice this in the spec sheets of the Genesis boxes too; basically the more drivers you combine, the less efficient the combination is overall. (IE, four compression drivers on a layered combiner *aren't* 6dB more efficient than one driver.)
 
My idea is to use two Eminence Alpha(lite) 6As in parallel for the midrange (low 100s through about 2kHz) and a fairly cheap compression driver above that. Sensitivity should be at least 100dB at 2.83V, though each individual box doesn't need a huge amount of power handling.

Here's a rough sketchup of what I am thinking
GoVfpg2.png


Outer dimensions are 20cm high x 60cm wide x 40cm deep (8x24x16 inches)

But I'm a bit more concerned how the directivity of a final unit is going to wind up.
My Paralines are similarly sized, 16.5 x 11.25 x 15 inches using Alpha 8s.
I did not port them, as LF in multiples is OK.
They work well in multiples, but the EQ required singly is quite different.

They can be used singly, but the vertical dispersion being very narrow at HF going to near omni at 1K is not very useful.

Art
 
Some thoughts:

1) What's the application?
2) Personally, I am not as concerned about the dips and peaks in the frequency response as most people are. For instance, if you make a call on your phone, the frequency response is far from flat. But the sound is not offensive. Now try making the phone call on a VOIP service like Skype. Even the smallest 'blip', a fraction of a second, is audible and offensive. That's how our hearing mechanism works; it's very sensitive to timing.
3) Pattern flip makes horns sound 'wonky' it's true. But there's an easy solution to pattern flip with the VTC boxes; just array them like they're supposed to be used. In your application that may be an acceptable solution.
4) Instead of investing a lot of time with Akabak and Hornresp, I'd simply download the dimensions of the VTC boxes and start from there. They've published all of the data in CLF format, so you can see *exactly* what the frequency response and directivity is like.


Thanks for your thoughts, John.

1) Basically the thought is a modular, mobile PA system. The basic item I want from a single box is a "boombox", high efficiency, no bass, battery powerable - but ideally I'd like to be able to combine a few to make a full-fangled PA setup (stack of 5 perhaps).

The whole project grew out of optimizing my previous boombox setup, so this is still the primary goal. However I discovered that in order to get >100dB sensitivity, I will pretty much automatically get about 100W worth of power handling, so might as well allow the system to make use of that if need be.

2) I have a psychoacoustics background, so I know a fair amount about the human hearing system - unfortunately my knowledge is rather lacking in terms of electro-acoustics and just general acoustics in air and horns. I've listened to some pretty wavy sounding FR speakers and still enjoyed the music, and particularly for this sort of application, hifi takes the backseat to SPL. But I'm a perfectionist like that, I'd still like it to be accurate (on paper at least).

3) well unfortunately my box is a bit smaller in each dimension than the VTC box, most noticeably in height, because it uses 1 tweeter instead of 2, and therefore 6.5" instead of 10" woofers. I feel that this is just about the biggest I can go for a single box, and when not arrayed, the 20cm (8 inch) vertical height is simply quite worryingly small for a horn mouth. I guess I might just have to build and measure and see how it turns out, practice is often the better arbiter than theory, in Germany there is a saying "Probieren geht über Studieren," which means to try is better than to study.

4) Good idea, I'll definitely look into that. Some conclusions can definitely be drawn even from other dimensions, particularly accounting for the respective wavelengths


Having said all that, the main reason I like Paralines is that it allows you to package the horn in some very odd shapes. For instance, under the dash of a car, or have the horn fire one direction while the driver points another. In your app, a plain ol' Synergy horn might give you better results. I don't see that a Paraline would offer any efficiency advantage, as a fair amount of output is lost on the high end due to imperfect combining of the high frequencies.

Indeed, I don't expect an efficiency advantage, in fact I believe that my woofers and their corresponding low volume are likely to be the limiting factor in total efficiency. Rather what I expect from the paraline is idealy a narrow vertical dispersion pattern, so everything doesn't go up into the air / bounce into the ground. The logic of the coherent line source wave pattern emanating from the front of the paraline (the original style, not your square version) appeals to me (though again, my knowledge of wave acoustics outside of free field point sources is unfortunately rather limited, so I don't know how accurate this picture is), however I don't know how long that vertical line source wave front can hold up once it encounters the diffractive horn mouth edge. As I mention, my experiments towards this in the hornresp wavefront simulator have been rather inconclusive.

You'll notice this in the spec sheets of the Genesis boxes too; basically the more drivers you combine, the less efficient the combination is overall. (IE, four compression drivers on a layered combiner *aren't* 6dB more efficient than one driver.)

This is obvious, you're never going to get more than 1 Watt out of 1 Watt in, no matter how many drivers you combine. so 1 Watt electrical power will never give you more than 109dB @ 1 meter for a true point source radiating into 4 pi. But a great way to drive this figure up is to reduce the angle it's radiating into.

Half space allows 112dB from 1 Watt already, but if you're able to get down to a radiation angle of 90 x 25 degrees, this (by my quick arithmetic using a rectangular approximation of sphere surface area around 0.2pi steradians) should allow you to get 122dB from 1 Watt. Twice the loudness without breaking any physical laws of engery conservation.

I hope I've brought the point across why I am concerned about losing vertical directivity, for the pursuit of sensitivity / efficiency it is simply wasted energy. Oh and potential bothersome for neighbors. Wow I've blabbered on quite a bit now, I'll stop.

:scratch::drink:

One other question though, the B&C DE 10-8 is the cheapest compression driver I can find, but it only claims 40W power handling from 2.5kHz up, so probably about half that if I crossover at 2kHz or slightly above, which at the nominal efficiency of 107dB would still allow for 120dB max, but I feel like that might be cutting it a bit close. Agreed? As an alternative the BMS 4538-8 seems like quite a good bargain, not much more expensive, made here in Germany and already offering a nominal 114dB sensitivity in a 90x75 horn (though the broadband sensitivity is less, judging from the graphs), it will handle 60W AES from 1.9kHz up, which should definitely keep up with the little Alpha 6's. Anybody have any experience with this driver? Forum search only turns up one hit. BMS in general seem to have a good name, but I guess that is quite an entry level product.
 
My Paralines are similarly sized, 16.5 x 11.25 x 15 inches using Alpha 8s.
I did not port them, as LF in multiples is OK.
They work well in multiples, but the EQ required singly is quite different.

They can be used singly, but the vertical dispersion being very narrow at HF going to near omni at 1K is not very useful.

Art

Hi Art,
thanks for that. I thought I remembered you having a fairly similar setup, but since it's nearly a year since I looked into it I couldn't remember any details nor did I have any info bookmarked. I've learned from that - documenting my progress better this time to avoid doing things more than once!

Anyway, I suppose any object that small is going to get omni at a relatively high frequency, there is nothing I can do in terms of design to change that. Too bad, but still, an inherent limitation I will have to live with, and design around. My thought would have been to plug the ports for multiples, but I can play around with that.

I guess I'll just order the parts for one and do some measurements and see what I come up with. About 200 euros in parts and material for the whole shebang seems reasonable enough, though it does add up once you start stacking them in the dozens.
 
James,
I don't see where the reference to going omni at high frequencies is coming from. I would think that it would be just the opposite and the omni directivity would be at low frequency and that at higher frequency at least with a single unit you would just see a pattern flip at the dimension of your vertical height where the upper frequencies would flip radiation pattern?
 
Hi Art,
thanks for that. I thought I remembered you having a fairly similar setup, but since it's nearly a year since I looked into it I couldn't remember any details nor did I have any info bookmarked.


I guess I'll just order the parts for one and do some measurements and see what I come up with. About 200 euros in parts and material for the whole shebang seems reasonable enough, though it does add up once you start stacking them in the dozens.
The Paraline with 2x8" is working well for PA use in multiples for me.
That said, I am using fourteen 8" drivers and five EVDH1AMT high frequency drivers per side, to get the equivalent line length using 6" drivers would require even more drivers.

For a single (or a couple per side) cabinet use the cabinet design pictured below works much better for a pair of Alpha 6", is easier to build, and uses the inexpensive little Eminence APT for the tweeter. This arrangement allows around a 3500 Hz acoustic crossover, the 6" cover the voice fundamental range, the tweeter is really only covering harmonics, so has no problem "keeping up".
A very even 90 degree horizontal pattern, and no vertical pattern flip in the midrange like a single Paraline has.
Passive crossover has a surprisingly low parts count, yet has good phase response through the crossover region.

It is the same physical size as my Paraline PA cabinets, but oriented completely different, 26.5 tall x 15 wide and 11.25 inches deep.

Sounds good right up to around 125 dB at one meter, and is over 100 dB with one watt midband.
Needs a fair amount of EQ, but not nearly as much as a single Paraline.
I have used a pair per side of these 2x6T stacked horn to horn with a sub below for outdoor gigs with around 1000 people, everybody was happy.

Art
 

Attachments

  • 2x6T.jpg
    2x6T.jpg
    94.9 KB · Views: 1,165
  • 2x6T Covered.jpg
    2x6T Covered.jpg
    100.6 KB · Views: 1,036
  • 2x6T xover.jpg
    2x6T xover.jpg
    102.6 KB · Views: 1,011
  • 2x6T F&P R.png
    2x6T F&P R.png
    164.8 KB · Views: 998
Last edited:
James,
I don't see where the reference to going omni at high frequencies is coming from. I would think that it would be just the opposite and the omni directivity would be at low frequency and that at higher frequency at least with a single unit you would just see a pattern flip at the dimension of your vertical height where the upper frequencies would flip radiation pattern?
James wrote:
"I suppose any object that small is going to get omni at a relatively high frequency, there is nothing I can do in terms of design to change that."

The "relatively high frequency" that the single Paraline "goes omni" in the vertical pattern is around 1000 Hz, it then narrows progressively at high frequencies.

The design I posted in #512 avoids pattern flip in that fairly critical frequency range.
 
img-placeholder-single-product-banner.png


On this forum, Tom Danley has given us some information on some interesting new ways to fold a horn. These new horns include the VTC Paraline, which combines two compression drivers on one horn, and the Jericho J4, which uses what he calls a 'layered combiner' to load 64 compression drivers on one horn.

Unity Paraline - YouTube

I decided to build one this week.

Once you figure out how they work, they're a lot of fun to build. They're waaaaaaaaaay easier to assemble than a Synergy horn, that's for sure. I built this one in under two hours.

The device that I built is neither a Paraline nor a layered a combiner. For the most part, it's a simplification of the Paraline. (I could build a *real* Paraline, but I didn't want to for this project, I wanted to do a 'proof of concept' first.)

As usual, my projects are completely kludgey and rushed and crude, but you get the general idea.

In the video, I think you'll notice a few things:

#1 - Although my compression driver appears to be firing through a block of wood, I have extension out to at least 18khz
#2 - Have you ever seen a horn that was less than four centimeters deep?
#3 - Have you ever seen a Unity horn that was smaller than a 8" woofer?

Hope you enjoy the vid.

j3.jpg


Danley posted a video of a new Jericho horn called a "J5 Caleb" on Youtube and Facebook this week. No info on the speaker on the website.

In one of their videos I noticed this monster horn being assembled in their factory. That's a lotta drivers!

(I don't know which Jericho horn that is; may be a J5, may not.)
 
Danley posted a video of a new Jericho horn called a "J5 Caleb" on Youtube and Facebook this week. No info on the speaker on the website.

In one of their videos I noticed this monster horn being assembled in their factory. That's a lotta drivers!
The Caleb appears to be two part, 54 drivers in each.

Mike Heddon stated dispersion of 40 x 15 degree, Ivan Beaver says driven by an iPod headphone output it is audible to past 150 feet.
 

Attachments

  • Caleb.png
    Caleb.png
    263.6 KB · Views: 477
Hi Art, Patrick, all
Art that’s right in fact as a joke we played it with an iphone at each demo at the Infocomm show earlier this month. Sensitivity is your friend haha.
Mike had those who wanted come up and step into the horn and move around to demonstrate even at this driver count, you can’t hear more than one source of sound floating in front of you or any change as you move around.

If you have Facebook, the company page has a number of videos put up today taken at 195 feet and 600 feet with the Matterhorn for low end.

https://www.facebook.com/DanleySoundLabs?ref=ts

Best,
Tom
 
Last edited:
Hi Art, Patrick, all
Art that’s right in fact as a joke we played it with an iphone at each demo at the Infocomm show earlier this month. Sensitivity is your friend haha.
Mike had those who wanted come up and step into the horn and move around to demonstrate even at this driver count, you can’t hear more than one source of sound floating in front of you or any change as you move around.

If you have Facebook, the company page has a number of videos put up today taken at 195 feet and 600 feet with the Matterhorn for low end.

https://www.facebook.com/DanleySoundLabs?ref=ts

Best,
Tom

I have two questions and an idea.

First, an idea -
Have you ever considered offering your Jericho speakers as a modular unit? For instance, it looks like you could subdivide many of the Jericho horns into four or even sixteen pieces. This would allow small crews to assemble a Jericho horn without a crane. This would work especially well with the injection molding that's used on some of your Synergy horns. Basically the horn could 'lock' into place.

And two questions:
1) I know that Danley doesn't do line arrays. So I'm guessing that the wavefront that exits the Jericho horns is curved both vertically and horizontally?
2) And if that's the case, why not just use the Paraline, instead of the layered combiner? A vertical array of Paralines can create a curved wavefront.
pie8.gif

To do this trick, each of the vertically arrayed paralines would be responsible for projecting slices of a pie, kind of like the pic above. Except only half the pie.
 
Hi Kessito
Your right they aren’t like normal. They don’t need to go up as high and do need to go down lower than in the smaller synergy horns. The acoustic load presented to the hf drivers make that possible. The mid drivers are in a sealed back volume though, but it is larger than a sealed back. You can’t break the thumb rules or your thumbs will break.

Hi Patrick
Well the J-5 is made in 4 sections, two flare extenders on the front and two driver sections on the rear.
There are locking over center clamps that attach the front to the back and gravity and interlocking feet takes car of the stack. That being said, I have been asked to slice it up further top to bottom. “A problem” is that this might make people inclined to make it taller or shorter than it is and to avoid pattern flip, the H and V angles govern the height and width. The act alike / sound like they do because they act like a single simple radiating source and goofy non-CD radiation is something to avoid. To the degree one can make it a simple wideband CD source, it will sound the same at 10 or 300 feet .

Ans1 yes there is a wave front curvature at “throat” of the final horn in both H and V. The throat is too wide (H) for free radiation into even 40 degrees up high, I don’t want a diffraction style horn so what enters the final part is already curved like it came from a point further behind.

Ans2 yes one can do this, which way makes the most sense depends on how much power density one needs. It’s too bad you couldn’t go to that Infocomm show because in addition to the iphone headphone jack powering the J5, we demonstrated a 5 foot tall 8 inch wide “column speaker” that is actually 130 degree H by 15 degree V horn. It is made of a stack of 8 Paralines driving a final flare. Each adjusted in size to curve the ends for the vertical part and each with the driver offset and “EYE” corrected for being off axis so that what exits is a continuous wave front, even with pink noise, it “sounds like” a single source with a very high directivity in the vertical plane.
This would be like your drawing, say facing to the right and a narrow working angle, divided into 8 sections.
Best,
Tom