CCBT (cheap constant beamwidth technology)

Status
Not open for further replies.
Jump to Latest
#2
xmQKq43.jpg


Once you get past the math, the CBT is simple:

1) curving the array controls the beamwidth. For instance, if your CBT covers 45 degrees, then your beamwidth is going to be approximately 45 degrees
2) shading the array improves the polar response

The first thing that I noticed about the shading is that a third of the array is barely generating any output whatsoever. Thirty percent of the drivers are generating less than 10% of the arrays output.

So... What if we just remove them? What happens?
 
#3
6xgvmin.png

Here's a ten element CBT array with Legendre shading.
The array is 25" tall and 5" deep. (64cm x 12.7cm)
The weighting on the elements is the same as the CBT in post #1

An externally hosted image should be here but it was not working when we last tested it.

Here's the exact same array, except I omitted 20% of the elements, the elements at the ends of the array that are getting a small fraction of the output. I'll call this thing "CCBT #1"

khJ9kEf.png

Here's the response of a ten element CBT array with Legendre shading at 1000Hz

An externally hosted image should be here but it was not working when we last tested it.

Here's the response of "CCBT #1", an eight element CBT with 2 elements omitted
 
#4
xyzOt4y.png

Here's the response of a ten element CBT array with Legendre shading at 2000Hz

Mej39eC.png

Here's the 2000Hz response of "CCBT #1", an eight element CBT with 2 elements omitted

yVQsDe7.png

Here's the response of a ten element CBT array with Legendre shading at 3000Hz

iAPPHdM.png

Here's the 3000Hz response of "CCBT #1", an eight element CBT with 2 elements omitted


pDrXgTB.png

Here's the response of a ten element CBT array with Legendre shading at 6000Hz

DL7VUyi.png

Here's the 6000Hz response of "CCBT #1", an eight element CBT with 2 elements omitted
 
#5
Some random observations:

1) Clearly the center-to-center spacing is critical. Neither design looks very good above 6000Hz. The center to center spacing is 2.5", which sets a (theoretical) upper limit of 5400Hz. (Yes, the drawing is done to scale.)

2) I'll concede that the response off-axis is a LITTLE bit better with the ten element array, but I'm not sure if it's good enough to justify a price tag that's 25% higher than the eight-element variant. Note that the beamwidth of the two speakers is very very similar, because the drivers at the edge of the "real" CBT are barely generating any output at all. YMMV.

Next up, I'll see if log shading will save some money over conventional shading.
 
#6
Hi John,

you cannot just omit the shading without large penalties, its the equivalent of a horn without roundovers; ugly diffraction and none-constant directivity response.
Please save yourself the trouble, it will not work well.
 
#7
kessito said:
Hi John,

you cannot just omit the shading without large penalties, its the equivalent of a horn without roundovers; ugly diffraction and none-constant directivity response.
Please save yourself the trouble, it will not work well.
Click to expand...

These are pretty darn close:

iAPPHdM.png


yVQsDe7.png


Here's why:

I'm not ELIMINATING the shading; I'm just eliminating a fraction of the array.

80% of the array is still there.

In a "conventional" CBT array with Legendre shading, we have the following levels for the ten elements:

-12, -9, -6, -6, -3, -3, 0, 0, 0, 0

In my "CCBT" array, I have the following levels for the eight elements:

-6, -6, -3, -3, 0, 0, 0, 0


I won't deny that the CBT is a tiny bit smoother. The question is:

Is that tiny bit of extra smoothness worth 25% more cost?



On a side note, I have a hunch that I could improve the polars of the "CCBT" if I used a better roundover at the top. All of my latest experiments are pointing to the fact that the shape of the enclosure is every bit as important as driver spacing and intensity, and you want an enclosure that basically looks like a bar of soap. NO SHARP EDGES
 
Status
Not open for further replies.
Top Bottom