The CBT is finished! I don't have it installed in the theatre yet, I'm in San Jose, CA at the moment but if all goes to plan I will be home this weekend and have it installed by early next week.
It performs pretty much exactly as the math predicts. We made measurements on axis to 70 deg off axis in 5 deg increments but I'm posting 10 deg increments to keep the graph less crowded. The lowest curve is 70 deg off axis, just count lines up from there to see the result of 60 deg off, 50... etc. By the time you get to 30 deg off axis, the lines are very nearly aligned with the on axis response. As you can see, the off axis response drops off very uniform within the CBT's range. Starting about 240 Hz, it begins to loose control of directivity and by about 160 Hz, it's no longer acting like a CBT. Because it's open baffle however, it still maintains some control over the low frequency off axis response.
We also made an on axis measurement with and without a rear wall reflection. To do this, we set up a 4 x 8 sheet of MDF behind it at the position where the rear wall will be in the theatre. In this graph the two red curves are with the wall in place, the blue curve is without the wall. The out of control wiggly red line is with the wall bare. The other red line is with the rear wall reflection absorbed by Owens Corning 703 fiber panels. These are 2" thick. We had two layers over the entire MDF board, three layers on the bottom half of the board and then additional 2 x 4 sheets right behind the center of the speaker. We also had some 1" sheets and the thickest part was in total 15" thick. I've just ordered twenty four more 2" sheets to add to this once it's in the theatre.
We wrote all over my driveway to do this test, you can see the marks where we drew out the test points for the 5 deg increments.
We also have a 1" layer of the fiber directly behind the twenty 6" Dayton Reference drivers and the fifty Aura 2" drivers. There is no fiber behind the one hundred thirty two Tang Band tweeters because those are closed back anyway.
Given the direcivity control this speaker exhibits I think the dialog quality should be unparalleled. It's very important to absorb that rear wave however. The nice thing is that you know exactly where to put the fiber... the thing is pretty much like a magnifying glass focusing that rear wave toward the center point of the 120 deg arc of the speaker. This arc is on a 49" radius which will be the distance to the rear wall. I'm going to have a lot more absorption there than we used in the outdoor test.
Once I get it installed I'll post some photos of that. Maybe I'll eventually get around to updating my web site with more photos and details.
All seven channels of the Emotiva UPA-7 amplifier are used. Three channels for the 6" drivers and two each for the 2" and 0.5". The Behringer DCX 2496 sets the - 3 dB increments for the 2nd and 3rd driver banks on the 6" drivers. Resistors set the attenuation on the 4th and 5th banks. There is also a 0.5 dB attenuation between the crossover and the amplifier on the center bank of tweeters. It's too bad the Behringer didn't have seven outputs, that would have eliminated the need for this external 0.5 dB attenuation. It was of course measured using Arta, the mic is the Earthworks M30.
If you are not familiar with CBT (Constant Beam width Transducer) arrays google Don Keele and read his papers on the topic.
mk
It performs pretty much exactly as the math predicts. We made measurements on axis to 70 deg off axis in 5 deg increments but I'm posting 10 deg increments to keep the graph less crowded. The lowest curve is 70 deg off axis, just count lines up from there to see the result of 60 deg off, 50... etc. By the time you get to 30 deg off axis, the lines are very nearly aligned with the on axis response. As you can see, the off axis response drops off very uniform within the CBT's range. Starting about 240 Hz, it begins to loose control of directivity and by about 160 Hz, it's no longer acting like a CBT. Because it's open baffle however, it still maintains some control over the low frequency off axis response.
We also made an on axis measurement with and without a rear wall reflection. To do this, we set up a 4 x 8 sheet of MDF behind it at the position where the rear wall will be in the theatre. In this graph the two red curves are with the wall in place, the blue curve is without the wall. The out of control wiggly red line is with the wall bare. The other red line is with the rear wall reflection absorbed by Owens Corning 703 fiber panels. These are 2" thick. We had two layers over the entire MDF board, three layers on the bottom half of the board and then additional 2 x 4 sheets right behind the center of the speaker. We also had some 1" sheets and the thickest part was in total 15" thick. I've just ordered twenty four more 2" sheets to add to this once it's in the theatre.
We wrote all over my driveway to do this test, you can see the marks where we drew out the test points for the 5 deg increments.
We also have a 1" layer of the fiber directly behind the twenty 6" Dayton Reference drivers and the fifty Aura 2" drivers. There is no fiber behind the one hundred thirty two Tang Band tweeters because those are closed back anyway.
Given the direcivity control this speaker exhibits I think the dialog quality should be unparalleled. It's very important to absorb that rear wave however. The nice thing is that you know exactly where to put the fiber... the thing is pretty much like a magnifying glass focusing that rear wave toward the center point of the 120 deg arc of the speaker. This arc is on a 49" radius which will be the distance to the rear wall. I'm going to have a lot more absorption there than we used in the outdoor test.
Once I get it installed I'll post some photos of that. Maybe I'll eventually get around to updating my web site with more photos and details.
All seven channels of the Emotiva UPA-7 amplifier are used. Three channels for the 6" drivers and two each for the 2" and 0.5". The Behringer DCX 2496 sets the - 3 dB increments for the 2nd and 3rd driver banks on the 6" drivers. Resistors set the attenuation on the 4th and 5th banks. There is also a 0.5 dB attenuation between the crossover and the amplifier on the center bank of tweeters. It's too bad the Behringer didn't have seven outputs, that would have eliminated the need for this external 0.5 dB attenuation. It was of course measured using Arta, the mic is the Earthworks M30.
If you are not familiar with CBT (Constant Beam width Transducer) arrays google Don Keele and read his papers on the topic.
mk
,I bet it sounds as good as it looks..
I just had to do it for a pair of baffles with 16 tweeters each, and that was already tedious.
