Constant Beamwidth Transducer (CBT) Speakers

[Head_Unit]

if anyone lives near Silver Spring MD I would be happy to give them the CBT36K

No I am in L.A., ya think DoorDash would pick up and deliver?!?

Roger Russell's IDS-25 column-array design

Huh...I swear that guy made a sales call when I worked at an automotive company. Looking at the site, what is special about it versus any other "umpteen driver" column? I've thought about such designs however due to the large number of drivers they end up yeah not cheap. And as a loudspeaker engineer I dislike series wiring* which becomes almost unavoidable. OTOH if you did an 8' column you don't need the CBT, it IS a CBT or at least plane wave launcher. Ideally you'd bury a line of those flush into the wall, eliminate those first reflections.

*hmmm let's say 3" drivers so 4 per foot = 32 drivers. At 8Ω/32=0.25 ohms. I suppose broken into chunks and driven by a 4-channel or one of those multichannel amps might work...groups of 4 would yield 2Ω and really any good sized amps should be fine so long as you avoid heavy clipping. Or use stacks of mini-Class D amps.

 

[somebodyelse]

*hmmm let's say 3" drivers so 4 per foot = 32 drivers. At 8Ω/32=0.25 ohms. I suppose broken into chunks and driven by a 4-channel or one of those multichannel amps might work...groups of 4 would yield 2Ω and really any good sized amps should be fine so long as you avoid heavy clipping. Or use stacks of mini-Class D amps.

You'd just use a series/parallel combination to get them to a sensible impedance - 8 groups of 4 in series, or vice versa.

 

[rdenney]

No I am in L.A., ya think DoorDash would pick up and deliver?!?

Huh...I swear that guy made a sales call when I worked at an automotive company. Looking at the site, what is special about it versus any other "umpteen driver" column? I've thought about such designs however due to the large number of drivers they end up yeah not cheap. And as a loudspeaker engineer I dislike series wiring* which becomes almost unavoidable. OTOH if you did an 8' column you don't need the CBT, it IS a CBT or at least plane wave launcher. Ideally you'd bury a line of those flush into the wall, eliminate those first reflections.

*hmmm let's say 3" drivers so 4 per foot = 32 drivers. At 8Ω/32=0.25 ohms. I suppose broken into chunks and driven by a 4-channel or one of those multichannel amps might work...groups of 4 would yield 2Ω and really any good sized amps should be fine so long as you avoid heavy clipping. Or use stacks of mini-Class D amps.

Roger Russell was the head of the speaker design group at McIntosh back before they were a fetish company, and it was he that established a proper measurement protocol for them, including building and using a large anechoic chamber. A lot of water has gone under the bridge since those days, of course. The IDS-25 was more recent.

He has several AES papers on the topic of vertical arrays of full-range drivers. His idea was that it eliminated vertical pattern effects in addition to other salutary effects. I didn’t dig into it that much, and I don’t have access to those papers, but I would reject his notions only after careful consideration.

It’s clear he was not a web designer, however

I included it because it’s part of the literature for vertical column arrays like the JBL CBT. Take it for what it’s worth.

Rick “doesn’t know how they were wired” Denney

 

[Head_Unit]

You'd just use a series/parallel combination to get them to a sensible impedance - 8 groups of 4 in series, or vice versa.

That's what usually happens, but I dislike series wiring since the impedance curves never quite match so the power division is not correct. Yeah 1st world problems but I'm an engineer so OCD about stuff like that.

 

[Bjorn]

What I am saying is that because of the low crossover frequency the tweeters are driven into a non-linear excursion by the energy in the low end of crossover passband for them. The tweeter are actually the same as some drivers used in laptop computers.

Do you have the older cabinet with a separation between midwoofers and tweeters or the newer cabinet where there's no separation between them and not much bracing?

What you're describing shouldn't exist with a crossover at 1 kHz and never did with the CBT36 pair I had (older and proper cabinet). I even lowered the crossover to 700-750 Hz area without any issues.

 

[jaffar34]

Hi everyone
My name is Claude ; i'm French and CBT addict . I'm probably the only one in France ( what a fool ! ) .

I decided to make my own when Jim GRIFFIN presented his prototype on another forum .
And i made this

It's a groundplane CBT with 24 2.5 " SB acoustics SB65 . But unlike Griffin or the Dayton Epique , the wiring is divided into 6 banks .
I've tried different subwoofers and the best , the one that goes best with the array is a DOS subwoofer ; the best solution for a perfect blend .
I 'am currently working on an improvement , an upgrade , in the form of a tweeter line . It will look like this

Th line will include 60 tweeters SB acoustics SB14ST per channel spaced 26 mm on a steel frame .
I can give you more details if you want .
Comprehensive set of measurements will be done . My goal is to smooth the response curve up to 15 kHz , and improve high frequency definition .

Another thing : it's possible to "buy" a ground plane CBT with the Martin Audio O'Line . But the price ....

 

[Paffi]

Another thing : it's possible to "buy" a ground plane CBT with the Martin Audio O'Line . But the price ....

I've the impression CBT is especially expensive in Europe the cbt 70j costs 1.900$/pair in the US and in Germany it's 3.200€, so around 3.800$, that's double. That premium is insane, feels ripped off.

 

[jaffar34]

Yes . If you need to multiply the number of speakers it affects the cost . I had to keep an eye for sales for my projects . The CBT36K used tweeters for one euro .
I paid quite 20 euros for each SB14 ...

 

[jaffar34]

This is what you never see about line arrays or groundplane CBT : a quasi-anechoic spl response curve unsmoothed ...

You can now understand why i want a line of tiny tweeters .

 

[Newman]

@jaffar34 can you identify the specific loudspeaker to which that measurement applies, please?

 

[jaffar34]

Hi Newman
This curve refers to my groundplane CBT array , photo above : 24 speakers SBacoustics SB65 , ctc 65 mm before any signal processing .
By comparison this is the spl response curve i measured of a single unit .

 

[Bjorn]

Fullrange drivers in a CBT is an obvious compromise for several frequencies, including the highs.

 

[jaffar34]

Obvious is the word .
I have measured several tiny speakers from aurasound , visaton , tang band , peerless , dayton and apart from the tang band 13-1761 , none are tweeters . They all are mediums .
I was fascinated by the NSW1 from aurasound but it's not able to make trebles .
The sb acoustics 65 was the best compromise for an affordable experience . I'm not disappointed .

@ Bjorn : i'd bet on the visaton BF32 for your new project . Am i right ? I would really like to know what your measurements look like .

 

[jaffar34]

CBT tweeters currently being assembled

 

[lilin]

Hi everyone,
I’ve recently been fascinated by CBT and am trying to grasp its technical nuances. However, I’ve hit a conceptual wall that I’m hoping you can help me break through.

According to the JBL CBT Series Setup Guide (Figure 1), these speakers cannot be stacked to extend the array in the same way as traditional line arrays. Their radiation pattern seems to behave more like a point source with a highly controlled vertical beamwidth. This observation is also supported by simulations in the JBL CBT Calculator (Figure 2).

Initially, I assumed CBT was simply a line array with superior side-lobe suppression (as shown in Figure 3, left). But now, it appears the technology mimics a point source located at the cabinet's acoustic center, radiating within a defined vertical window (Figure 3, right).

While both patterns mitigate floor/ceiling reflections, I’m concerned about domestic applications: If the "point source" model is correct, would the short listening distances in a typical home lead to an overly narrow vertical "sweet spot"? Would the vertical coverage be insufficient for listeners who aren't perfectly aligned with the acoustic axis?

 

[Paffi]

@lilin like you, I fell in love with the cbt and I had asked the same question in the jbl-thread (https://www.audiosciencereview.com/...w-constant-beam-transducer.21469/post-2570569).

There was no answer yet. But in that thread I found this helpful post:

JBL CBT 70J-1 Review (Constant Beam Transducer)

@BenB I failed to address the issue you raised of side lobes. I did notice a significant one with the CBT-36 design. I think some of it may have been due to using auto EQ / correction with a DEQX unit trying to fill in a notch with the on-axis response.

www.audiosciencereview.com

In addition, the FAQs of JBL (https://jblpro.com/es/site_elements/cbt-frequently-asked-questions) tell us that, point 7:

"No, the aiming axis for all CBT models is straight in front of the speaker. The CBT’s get aimed
vertically by tilting them downward."

So I guess simply sitting in front (of the full 70cm), but not below or above (when facing straight in front), is ok. At least that is, what I'm experiencing with my JBLs at home.

My question is, if positioning them in reverse, so with the curving at the top instead not the bottom, will be an issue when you still sit directly in front of it and not having them above tilted downwards (where the curving makes sense to reduce the level for people standing nearer). With this the positioning on stands would be more robust and someone may use the extension module as a stand as well. Practically making a floorstander with that (originally the extension module is meant to be added at the top).

 

[rdenney]

Yes, the JBL CBT speakers create a pancake of sound. As mentioned before, I used them in a church to bypass the podium microphones and the first several rows of pews that didn’t need reinforcement. In a home, this would minimize floor and ceiling reflections but would also be more demanding of listening height on a smaller room.

Rick “wondered if they used a phased array concept” Denney

 

[somebodyelse]

The papers and presentations have a fairly detailed description of the way the beamforming works. It uses a combination of reduced amplitude and increased delay as the distance from the centre of the array increases. The papers show both the curved version like the Dayton CBT24 and CBT36 where the delay is provided by the distance provided by the curve, and a physical line like the JBL where the delay is provided electronically. The Dayton also uses the floor to mirror the array - the centre of the array is at the floor.

If you wanted to stack the arrays and have them behave as a single longer array, you would also have to change the shading function accordingly. That doesn't seem to be an option with those JBLs - it seems like they're treating them as separate arrays. It would be interesting to see a model of how those separate arrays interact.

 

[Jaxx1138]

I see a lot of people referring to the elimination of the floor bounce. The CBT's (Half CBT)shown don't e eliminate the ground bounce but utilize it, actually need it, for proper response. The acoustic center is on the floor in this case.

 

[Jaxx1138]

Hi everyone,
I’ve recently been fascinated by CBT and am trying to grasp its technical nuances. However, I’ve hit a conceptual wall that I’m hoping you can help me break through.

According to the JBL CBT Series Setup Guide (Figure 1), these speakers cannot be stacked to extend the array in the same way as traditional line arrays. Their radiation pattern seems to behave more like a point source with a highly controlled vertical beamwidth. This observation is also supported by simulations in the JBL CBT Calculator (Figure 2).

Initially, I assumed CBT was simply a line array with superior side-lobe suppression (as shown in Figure 3, left). But now, it appears the technology mimics a point source located at the cabinet's acoustic center, radiating within a defined vertical window (Figure 3, right).

While both patterns mitigate floor/ceiling reflections, I’m concerned about domestic applications: If the "point source" model is correct, would the short listening distances in a typical home lead to an overly narrow vertical "sweet spot"? Would the vertical coverage be insufficient for listeners who aren't perfectly aligned with the acoustic axis?

Their sweet spot is quite large vertically(about 20 degrees of the ends.) and horizontally(about 120 degrees ). You can build a full CBT array that doesn't use the floor as reinforcement. As long as the curve and shading effects are implemented correctly. I don't remember how much of the dB gain over distance is retained from the line source inverse square of 3dB vs 6dB from a true line source.