I'd been waiting until I was done building these speakers before I posted about my process, but I'm close enough I guess that I can start and leave a couple of placemarks for the final phases.
Anyway, I've been bitten by the DIY bug. I don't think it's fatal, but I think it's pretty much incurable. The latest flare-up is a desire to build a line array. There's just something intrinsically sexy about a big stack of drivers to me that I find/found irresistible.
I'm also accountant.
Why is that relevant? It means I'm.. chea.. err, frugal. I wanted to spend as little as possible. I set myself a budget of $300 for the drivers. After some consultations with someone much more experienced in these things, I settled on the Aura NS3 194 8E 3" wide range driver. If you take a look at their posted FR chart, you'll see that it's relatively flat down to ~200 and relatively flat up to ~ 7,000 or 8,000Hz. There are some general ups and downs, but I think those will be easy enough to tame with a little eq'ing. That being said, the upper octaves are a little too raggedy, so I will be employing a horn tweeter, a Fostex FT17H, to handle these frequencies. This tweeter is rated for 5khz to 50khz. Why it goes so high, I have no idea -- unless they were thinking I had bats that needed some musical entertainment. The driver is also highly efficient, so it should mate well with so many aura drivers coupling. If I were going to do this "right", I would use a ribbon tweeter since they have dispersal patterns similar to a stack o' drivers; however, to stay in budget, I used this driver. For reference, I'm thinking that my twelve Aura drivers will produce ~116db above 200Hz and ~104db at 80Hz at 15watts/1 meter. In other words, it'll be LOUD. I also think I should be able to EQ the system to be able to get a flat FR down to the F10 to F15 of the system as a whole without taking much of a sonic hit.
So, how to design a line array.
There is only one "paper" I've found on the internet that covers the "conventional" wisdom of line array design. Link.
However, someone I trust has told me that the theories proposed have been falling out of favor.
So, the thinking behind the design is this:
To get the cleanest response, create an enclosure that mimics an IB installation. Using WinISD, I determined the best response for the 12 Aura drivers and then doubled (or even triple) the volume of the enclosure. On top of that, I will pack the entire enclosure with R19 to further increase the "perceived volume" of the enclosure (I will use some batting to create a barrier between the R19 and the back side of the drivers).
To increase the coupling between the drivers as much as possible, I positioned the drivers as close together and I thought I could. As it stands, there is probably ~1/4" between each of the drivers flanges. If I could, I'd prefer these drivers were stacked right on top of each other.
Another issue addressed by the the enclosure design was the baffle width. The concern for me here is the BSC. My thinking was that I'd want the baffle to be as narrow as possible so that the BSC doesn't come into play in the lower octaves where these drivers are going to have the toughest time reproducing a flat FR, so, I went as narrow as I thought I could -- which turns out to be the thinking behind some recent designs I've seen (e.g., this one).
The next is to figure out the height. Here are the issues I had to contend with
Since I know what the volume is supposed to be and I have the height and width, that just leaves the depth as a plug.
So, in the end, the final dimensions came out to be something like 5x64x18 -- I have to admit I don't remember the exact dimensions, I'll measure them soon and post the final numbers. I also have a small base unit to add the height necessary to get the tweeter at the right height. Here is a preliminary "design" of what I expected the final product to be like.
As for the crossover, I'm going active with this project, so I don't have to design it -- just pick the crossover point. I'm going to be going for a crossover point as high as possible -- maybe 8khz.
JCD
Anyway, I've been bitten by the DIY bug. I don't think it's fatal, but I think it's pretty much incurable. The latest flare-up is a desire to build a line array. There's just something intrinsically sexy about a big stack of drivers to me that I find/found irresistible.
I'm also accountant.
Why is that relevant? It means I'm.. chea.. err, frugal. I wanted to spend as little as possible. I set myself a budget of $300 for the drivers. After some consultations with someone much more experienced in these things, I settled on the Aura NS3 194 8E 3" wide range driver. If you take a look at their posted FR chart, you'll see that it's relatively flat down to ~200 and relatively flat up to ~ 7,000 or 8,000Hz. There are some general ups and downs, but I think those will be easy enough to tame with a little eq'ing. That being said, the upper octaves are a little too raggedy, so I will be employing a horn tweeter, a Fostex FT17H, to handle these frequencies. This tweeter is rated for 5khz to 50khz. Why it goes so high, I have no idea -- unless they were thinking I had bats that needed some musical entertainment. The driver is also highly efficient, so it should mate well with so many aura drivers coupling. If I were going to do this "right", I would use a ribbon tweeter since they have dispersal patterns similar to a stack o' drivers; however, to stay in budget, I used this driver. For reference, I'm thinking that my twelve Aura drivers will produce ~116db above 200Hz and ~104db at 80Hz at 15watts/1 meter. In other words, it'll be LOUD. I also think I should be able to EQ the system to be able to get a flat FR down to the F10 to F15 of the system as a whole without taking much of a sonic hit.
So, how to design a line array.
There is only one "paper" I've found on the internet that covers the "conventional" wisdom of line array design. Link.
However, someone I trust has told me that the theories proposed have been falling out of favor.
So, the thinking behind the design is this:
To get the cleanest response, create an enclosure that mimics an IB installation. Using WinISD, I determined the best response for the 12 Aura drivers and then doubled (or even triple) the volume of the enclosure. On top of that, I will pack the entire enclosure with R19 to further increase the "perceived volume" of the enclosure (I will use some batting to create a barrier between the R19 and the back side of the drivers).
To increase the coupling between the drivers as much as possible, I positioned the drivers as close together and I thought I could. As it stands, there is probably ~1/4" between each of the drivers flanges. If I could, I'd prefer these drivers were stacked right on top of each other.
Another issue addressed by the the enclosure design was the baffle width. The concern for me here is the BSC. My thinking was that I'd want the baffle to be as narrow as possible so that the BSC doesn't come into play in the lower octaves where these drivers are going to have the toughest time reproducing a flat FR, so, I went as narrow as I thought I could -- which turns out to be the thinking behind some recent designs I've seen (e.g., this one).
The next is to figure out the height. Here are the issues I had to contend with
- the tweeter needed to be at my sitting ear level
- the whole unit had to be 6ft or less to fit in my garage -- its future home.
- since something this narrow is going to be unstable by itself, I wanted to attach some sort of bass unit.
Since I know what the volume is supposed to be and I have the height and width, that just leaves the depth as a plug.
So, in the end, the final dimensions came out to be something like 5x64x18 -- I have to admit I don't remember the exact dimensions, I'll measure them soon and post the final numbers. I also have a small base unit to add the height necessary to get the tweeter at the right height. Here is a preliminary "design" of what I expected the final product to be like.
As for the crossover, I'm going active with this project, so I don't have to design it -- just pick the crossover point. I'm going to be going for a crossover point as high as possible -- maybe 8khz.
JCD