Constant Beamwidth Transducer (CBT) Speakers

[Bjorn]

That's an irony. Considering there's a polar in the picture you posted

And there a bunch of polars in Don's papers:

One moment, please...

It's possible to do a CBT horn/waveguide speaker in order to get a narrow horizontal directivity. We've talked about it and also done some sketches. Personally I don't see much merit over a traditional horn speaker though.

 

[AdamG]

I’ll just leave this here. If you don’t recognize it then you may be too young. It was used in some type of marketing but I can’t remember exactly. Maybe “Keep on Trucking” but I don’t remember the product association. These type of speakers make me think of this image:

 

[Keith_W]

That's an irony. Considering there's a polar in the picture you posted

Sorry that was a slip of the keyboard. I wanted to say CEA 2034 instead of "polar" but I typed "polar" instead

It's possible to do a CBT horn/waveguide speaker in order to get a narrow horizontal directivity. We've talked about it and also done some sketches. Personally I don't see much merit over a traditional horn speaker though.

Why not? A traditional horn speaker has poor vertical directivity. Wouldn't something like this solve that?

 

[j_j]

Reviving this topic. This is a screen cap from Don Keele's video showing the directivity of his CBT design:

View attachment 364716

Whilst it's undoubtedly cool looking, I would prefer to look at a polar response to see what it is actually doing.

But it got me wondering whether it would be possible to combine a waveguide or a horn with a CBT design to improve the horizontal dispersion characteristics. Is this another one of my stupid ideas, or does it have some merit?

The answer is "yes" but width of transducers gets interesting. There are a few ways to go about it.

 

[j_j]

Why not? A traditional horn speaker has poor vertical directivity. Wouldn't something like this solve that?

If it's constant beamwidth across frequency, are you SURE it's a bad thing?

 

[Newman]

Hi j_j, since you are attending to this thread at the moment, can I just check with you: aren't all the special and sweet characteristics of this speaker, when standing on the floor as shown in photos, pertaining to the vertical plane?

...ie, the term 'constant beamwidth' in the name is actually vertical beamwidth, despite the fact that many lay readers coming across the term 'width' in its name will think it applies to the horizontal plane?

...and that the actual horizontal dispersion of a CBT speaker, when standing on the floor as shown in photos, is exactly like it would be for a single woofer and tweeter lying side-by-side, ie not good?

cheers

 

[Keith_W]

The answer is "yes" but width of transducers gets interesting. There are a few ways to go about it.

"Yes" it's one of my stupid ideas? I am assuming you meant it has some merit. I can see from the picture @Bjorn posted earlier that it looks pretty wide, not to mention a curve like that would be pretty difficult to make and way beyond my DIY ability. For this reason I am thinking of a linear array with DSP delays. Yes, I know it will require an ungodly number of DAC channels and amplifiers but I am imagining a system where I can turn it into a phased array at the push of a button. However that only controls vertical directivity problem, there is nothing I can do about horizontal directivity apart from waveguides or horns. The solution to that would be a submarine sonar like partial sphere with drivers pointing in all directions, but that is way too extreme. So maybe something even simpler to construct - a panel speaker with many drivers on it, each with digital delay and shading to create a virtual dome.

Yes I am aware that straight line arrays and curved line arrays radiate sound differently. But I like tinkering and to me, doing it with digital delays opens up the possibility of endless years of fun. But the idea of having to provide >200 DAC and amplifier channels is a bit too crazy for me. I am a hobbyist and not a naval sonar engineer.

Don's illustration does not say whether that straight line array had delays in it or not and what kind of shading pattern was used. I am guessing not, based on other illustrations I saw on his website.

If it's constant beamwidth across frequency, are you SURE it's a bad thing?

I honestly did not know it was a constant beamwidth across frequency. If it is, then horns or waveguides are not needed? I thought it would behave more like a conventional speaker with omni at the bottom and beams at the top. The illustrations I have seen so far (like the cut and paste from Don's website above) shows excellent smooth vertical directivity. Maybe there are spinoramas buried somewhere in this thread, I should go look for them.

 

[j_j]

...and that the actual horizontal dispersion of a CBT speaker, when standing on the floor as shown in photos, is exactly like it would be for a single woofer and tweeter lying side-by-side, ie not good?

cheers

Not quite. It's the dispersion of the driver that renders those frequencies, so it does not have to be bad at all.

 

[Curvature]

Hi j_j, since you are attending to this thread at the moment, can I just check with you: aren't all the special and sweet characteristics of this speaker, when standing on the floor as shown in photos, pertaining to the vertical plane?

...ie, the term 'constant beamwidth' in the name is actually vertical beamwidth, despite the fact that many lay readers coming across the term 'width' in its name will think it applies to the horizontal plane?

...and that the actual horizontal dispersion of a CBT speaker, when standing on the floor as shown in photos, is exactly like it would be for a single woofer and tweeter lying side-by-side, ie not good?

cheers

JBL CBTs, especially the CBT-1000, look good when looking at the EASE data they provide.

Measurements for speaker JBL CBT 1000

www.spinorama.org

I believe the vertical polar map refers only to the measurement reference point. As you move up and down the radiated sound will be consistent due to the number of drivers and their position.

 

[Bjorn]

Why not? A traditional horn speaker has poor vertical directivity. Wouldn't something like this solve that?

Does it look like the horn speaker in the below link has poor vertical directivity when measured over a reflective floor at different heights?

Vera Audio midrange horn - polar measurements

Vera Audio is developing a horn speaker system. It consist of a midrange horn, a midbass horn and requires a bass/subwoofer solution. It will be active and offered with a separate high quality DSP. The 2-horn system look like the following: I'll show some measurements of the midrange/top...

www.audiosciencereview.com

 

[Bjorn]

But it got me wondering whether it would be possible to combine a waveguide or a horn with a CBT design to improve the horizontal dispersion characteristics. Is this another one of my stupid ideas, or does it have some merit?

FIY: The traditional CBT speaker is quite constant horizontally. It's wide with 180° beamwidth. If by improving you mean making the narrow horizontally, there are several ways to do that. But the regular CBT speaker is overall uniform horizontally with few side lobes.

Note that the traditonal CBT is only wide at a limited height. That's exactly what the 3D polar balloon is showing us.

 

[Bjorn]

Hi j_j, since you are attending to this thread at the moment, can I just check with you: aren't all the special and sweet characteristics of this speaker, when standing on the floor as shown in photos, pertaining to the vertical plane?

...ie, the term 'constant beamwidth' in the name is actually vertical beamwidth, despite the fact that many lay readers coming across the term 'width' in its name will think it applies to the horizontal plane?

...and that the actual horizontal dispersion of a CBT speaker, when standing on the floor as shown in photos, is exactly like it would be for a single woofer and tweeter lying side-by-side, ie not good?

cheers

There exist both ground plane CBTs and "non ground plane" (free standing CBT) like the JBL CBT speakers. Two different designs.

Constant beamwidth transducer (CBT) speaks of both the horizontal and vertical directivity.

The horizontal directivity will depend on the drivers, placement of the drivers, cabinet, and the crossover. It's not a problem to have drivers side by side in regards to horizontal directivty if there are no lobing occuring.

 

[Newman]

Not quite. It's the dispersion of the driver that renders those frequencies, so it does not have to be bad at all.

JBL CBTs, especially the CBT-1000, look good when looking at the EASE data they provide.

Sure, I didn’t mean to say it’s impossible, but that the sort of demo unit that Keele was showing, tended to comprise either a single column of small full-range drive units which is likely to beam horizontally in the HF and approach omni horizontally in the LF, or a column of tweeters beside a column of small bass drivers, which will tend to lobe horizontally in a similar manner to a small 2-way speaker lying on its side. There is a good reason why 2-way home speakers are arrayed vertically.

Like JJ says, the horizontal dispersion will be whatever the drive units do. The special beam properties due to the array’s design are all about the vertical.

The reason I bring it up is because, for home hifi purposes, horizontal directivity issues are more critical than vertical to listener preference. The floorstanding CBT is unique in the less critical plane, and nothing unique in the more critical plane.

I sometimes wonder how they would go in the home environment turned sideways, sticking out from the room’s side walls.

cheers

 

[Bjorn]

There is no horizontal lobing with the CBT36. Spacing and crossover avoids that.

In most small rooms, the vertical directivity is generally more important than the horizontal and as shown in the studies of Bech. But a well designed CBT will not have horizontal lobing and a generally very uniform dispersion of 180° beam width. All speakers become omni in the bass in small rooms due to room modes.

Turning a CBT on the side would be beat the whole purpose of the design.

 

[Ze Frog]

It's an awesome concept, I brought 48 peerless 2.5" driver's to do a pair, but then realised they are really too large and magnet assembly is too big to allow the curvature of baffle with close spacing, was my earlier days and didn't really know what I was doing. Would love to make a pair one day though, the idea of how it creates a bubble of sound that can be controlled somewhat is fascinating. Some good literature and content on CBT can be found by looking into its conception as a military application for submarine stealth, super interesting stuff.

 

[Hayabusa]

It's an awesome concept, I brought 48 peerless 2.5" driver's to do a pair, but then realised they are really too large and magnet assembly is too big to allow the curvature of baffle with close spacing, was my earlier days and didn't really know what I was doing. Would love to make a pair one day though, the idea of how it creates a bubble of sound that can be controlled somewhat is fascinating. Some good literature and content on CBT can be found by looking into its conception as a military application for submarine stealth, super interesting stuff.

You can still do the curve delays with dsp and multiple channels

 

[Ze Frog]

You can still do the curve delays with dsp and multiple channels

Yeah, but there's still something pretty cool about doing so in the more analogue way. Plus these things look kinda cool, definitely a conversation starter if nothing else.

 

[Hayabusa]

Yeah, but there's still something pretty cool about doing so in the more analogue way. Plus these things look kinda cool, definitely a conversation starter if nothing else.

An other try: move odd/even speaker slightly from the center line so the magnet does not touch?

 

[markus]

All speakers become omni in the bass in small rooms due to room modes.

Not sure they become omni because of room modes but because of wavelength vs. source/room size. Anyway, it's just how historically most designs work. Omni down low, beaming higher up. But the question is, is this what we actually want? Are other directivity patterns more desirable for stereo reproduction? Maybe even the opposite behavior, i.e. narrow down low and wide higher up?

 

[Ze Frog]

An other try: move odd/even speaker slightly from the center line so the magnet does not touch?

Yeah, that could be a good idea.