I needed 48 tweeters to make a 65" line array, so I can up with a very simple design that I can make quickly. I use purchased diaphragms, some perforated sheet and some readily available 3mm x 10mm x 60mm magnets. I am putting the tweeters in groups of 4 to make assembly easier. The array will use electronic curvature with a 24-channel DSP/amp shown in the last photo.
I've only made one tweeter so far, but it looks like this design will work just fine.
Online metals has .106" thick perforated steel sheets with 51% open area. I used a Ryobi flooring saw with a 5-3/8" steel-cutting blade to slice up the sheets. That worked a lot better than I expected, without making a mess.
After a small amount of sanding the cut lines, I painted the cut plates and glued on the magnets with epoxy. The diaphragms were round, but I used a bench sander to allow closer spacing. The housing for the diaphragms measure about 2.7mm thick, so the 3mm magnets work great.
And then CAREFULLY put on the top grid. You need to watch your fingers because the pieces will try to snap together violently.
The prototype works, and I only have to make 23 more 🙂. Notice that this can be used as a dipole, which is what I intend to do for this line array. The midrange array will have 28 TC6 2" drivers on each side, along with a 12" ripole woofer for each array.
The DSP/amp is done, and all the channels work and can be individually controlled. However, it still needs more code to control the curvature and shading. The ripole woofers will have their own amps. There is more info on the DSP/amp at this link: http://www.audiodevelopers.com/2024/03/17/another-prototype-line-array-amp/
I've only made one tweeter so far, but it looks like this design will work just fine.
Online metals has .106" thick perforated steel sheets with 51% open area. I used a Ryobi flooring saw with a 5-3/8" steel-cutting blade to slice up the sheets. That worked a lot better than I expected, without making a mess.
After a small amount of sanding the cut lines, I painted the cut plates and glued on the magnets with epoxy. The diaphragms were round, but I used a bench sander to allow closer spacing. The housing for the diaphragms measure about 2.7mm thick, so the 3mm magnets work great.
And then CAREFULLY put on the top grid. You need to watch your fingers because the pieces will try to snap together violently.
The prototype works, and I only have to make 23 more 🙂. Notice that this can be used as a dipole, which is what I intend to do for this line array. The midrange array will have 28 TC6 2" drivers on each side, along with a 12" ripole woofer for each array.
The DSP/amp is done, and all the channels work and can be individually controlled. However, it still needs more code to control the curvature and shading. The ripole woofers will have their own amps. There is more info on the DSP/amp at this link: http://www.audiodevelopers.com/2024/03/17/another-prototype-line-array-amp/
Last edited:
I haven't measured these yet, so I don't know where to set the crossover frequencies. I know that the full array will go much lower than a single tweeter, but it will be a while before I can measure them.
There is probably a way to calculate the frequency response of multiple drivers in a vertical array from the single-driver response, but I don't know the formula. When I made the vertical array with 96 1/2" drivers, the response surprised me, but I haven't measured that one yet, either.
There is probably a way to calculate the frequency response of multiple drivers in a vertical array from the single-driver response, but I don't know the formula. When I made the vertical array with 96 1/2" drivers, the response surprised me, but I haven't measured that one yet, either.
Last edited:
Yep--there will be a total of 12 unique tweeter channels and 7 unique midrange channels, with symmetry about the centerline, so that it looks like 24 tweeter channels and 14 midrange channels. That's enough to implement electronic curvature with relatively low error, or at least, that is what I convinced myself of many years ago when I made the array in this article: http://www.audiodevelopers.com/10-case-study-3-a-line-array-with-dsp/. I am building this one to replace that older version, because it is too hard to make changes to the DSP on that older version. Once I get this new one working, I'm going to update that older one with the newer DSP and software.
BTW, these channel counts are just for one side. The other side for stereo has its own 24-channel DSP/amp, so there are a total of 48 unique channels available. Two of the channels for each side are dedicated for a woofer and a subwoofer. For this line array, there are three channels on each side that aren't used.
BTW, these channel counts are just for one side. The other side for stereo has its own 24-channel DSP/amp, so there are a total of 48 unique channels available. Two of the channels for each side are dedicated for a woofer and a subwoofer. For this line array, there are three channels on each side that aren't used.
Last edited:
I wish I could say nice things about those AMT diaphragms, but there are way too many failures and too much variation in the resistance. Out of 96 diaphragms, only 73 were good, in that their resistance was between 3.1 and 4 ohms. 9 of them measured over 8 ohms, and 4 of those were completely open.
I've got 4 containers with 6 groups of 4 tweeters, and I measured the DC resistance of each tweeter, and each group is labelled. If you look carefully at the labels, you can see the variation between the diaphragms.
I wrote to the vendor, and they told me to send a copy of the shipping label (long gone) and a video of the problem. Yeah, right...
I've been replacing the ones that are out of spec, but this is taking a lot more time and money than I expected.
I've got 4 containers with 6 groups of 4 tweeters, and I measured the DC resistance of each tweeter, and each group is labelled. If you look carefully at the labels, you can see the variation between the diaphragms.
I wrote to the vendor, and they told me to send a copy of the shipping label (long gone) and a video of the problem. Yeah, right...
I've been replacing the ones that are out of spec, but this is taking a lot more time and money than I expected.
The perforated steel cannot conduct enough magnetic flux, because the cross-sectional area is very small. The steel is saturated near the magnets. Have you done an efficiency test?
The steel is .105" thick with 51% open area. That's thicker than the steel in the commercial tweeters that use these diaphragms, with about the same open area. The fact that the magnet structure can hold a screwdriver equally well all along the gap proves that there is adequate cross-sectional area. I had initially tried using .06" thick steel to make cutting the sheets easier, but I wasn't satisfied with the magnet structure. I haven't measured the efficiency, but it's not going to be an issue since there are 48 tweeters connected to 12 amplifiers for each line array. I've tested them by using a 1.5V battery and they sound as loud as I expected for their size.