• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Little Big Line Array Build

Aijan

Member
Joined
Jul 13, 2021
Messages
46
Likes
159
After building quite a few subwoofers and 2-way speakers over the years, I was looking for something different, and line arrays fit the bill nicely. Line arrays have some intriguing characteristics such as little loss of volume as you get further away, low distortion, and very slim enclosures.

Designing one isn’t simple, though, since drivers interfere with each other, causing anomalies in both frequency response and sound dispersion. To minimize destructive interference and use it to our advantage, it’s vital to minimize driver-to-driver spacing.

When making line array speakers, some companies use a large number of 1" tweeters coupled with several small mid-bass drivers. Unfortunately, the distance between two 1" tweeters can’t be made much smaller than 40 mm, and they don’t really go much below 2 kHz without a waveguide anyway. On the other hand, while 2" full-range drivers result in a slightly wider driver-to-driver spacing of around 60 mm, this is a small price to pay given their ability to reproduce frequencies down to 150 Hz, which is particularly useful for a home theater surround speaker use case I have in mind.

I’ve settled on using eight 2" Lavoce FSN020.71F drivers, featuring neodymium magnets and 3/4" voice coils. I could have used more, but in a typical sized room, they should be more than enough.

layered_build_cut_from_mdf_with_cnc_router.jpg


Curves certainly look nice, but they aren’t easy to make. One method for creating a curved enclosure is to form it from layers cut from a sheet of wood. I’m not very fond of this technique, as a lot of material goes to waste, but in this case, the enclosure will be very slim, resulting in little waste. Overall, I’ve cut nine layers from 18 mm MDF on a CNC router. Some layers have inner lips to mount the speaker baffle and the rear panel, plus one in the middle for central support. I also applied 1/2" roundovers to the speaker cutouts.

little_big_line_array_baffle.jpg


Eight drivers mean a lot of unsightly screws, which led me to mount the drivers from the back. However, that requires either a removable rear panel or baffle (or both) to access the drivers in the future. Since space is tight, I decided to make both removable by using threaded inserts at the back of the baffle and long screws, inserted from the rear panel, to attach the front and rear together.

front-and-rear-panels_inside_view_with_crossover_board.jpg


There are four banks of drivers, each with two 8-ohm drivers connected in series. The banks are then connected in parallel, resulting in a nominal load of 4 ohms. However, two filters are necessary to flatten the frequency response for all drivers: a low-shelf cut and a notch. These bring the nominal impedance up to 6 ohms.

little_big_line_array_crossover_uses_frequency_shading.png


From top to bottom, drivers 1 and 8, 2 and 7, 3 and 6, and finally 4 and 5 form each bank. High frequencies are increasingly shaded by 2nd-order filters for each bank as they move further from the center, which helps reduce destructive interference and control vertical directivity.

little_big_line_array_with_stand_rear_view.jpg

The speaker was originally designed for wall mounting, but I made a quick stand from scrap wood, incorporating the wall mount, so that it could stand on its own.

line_array_on_subwoofer.jpg


So, how’s the sound? The speaker needs a subwoofer to fill in the missing low end, but once I’d dialed in the subwoofer integration, I was greeted with clean, smooth sound, particularly in the mid-range, and the speaker can go loud for sure. At 95 dB/1m, there was no noticeable distortion thanks to the eight drivers working in tandem. As for the controlled vertical directivity, you can tell the upper frequencies decrease in volume when your ears are not on the speaker’s center axis. The effect is subtle but quite noticeable, especially when you go increasingly off-axis.

little_big_line_array_on_axis_frequency_response.png


The frequency response is mostly within ±3 dB between 150 Hz and 17 kHz (measured outside, gated at 10 ms). The critical 1 to 5 kHz region, where the ear is most sensitive, is quite flat with only a ±1.5 dB variation. There are some response anomalies starting at 10 kHz since the 2" driver isn’t really suited to reproducing these frequencies well, and there are also some destructive effects from nearby drivers. Fortunately, there’s very little content at these frequencies.

little_big_line_array_horizontal_directivity.png


Horizontal directivity is wide at ±60 degrees nominal up until 5 kHz, but it starts to narrow somewhat after that.

little_big_line_array_vertical_directivity.png


Vertical directivity is near ruler flat at ±15 degrees nominal down to 1 kHz and gradually gets wider below that. This well-controlled and narrow directivity allows precise steering of sound, which is especially well suited for applications such as surround sound speakers in a home theater, where some listeners are closer to a speaker than others.

little_big_line_array_impedance_plot.png


The impedance and phase plot shows no major issues other than a few minor resonances. The bump centered at around 2.3 kHz is due to the notch filter that flattens the response.

Finally, a complete build video is available below. Please let me know what you think.

 
Top notch build, congratulations!
Reminds me of one of my earlier projects:

This one did not need a sub though :)
 
Very cool build. Looks awesome. But looks a bit bumpy up top eh? I would assume it isn't really audible though. Congrats!
 
Top notch build, congratulations!
Reminds me of one of my earlier projects:

This one did not need a sub though :)
Impressive build. Thanks for sharing!
 
Very cool build. Looks awesome. But looks a bit bumpy up top eh? I would assume it isn't really audible though. Congrats!
Thanks! The bumpiness above 10 kHz is clearly audible in frequency sweeps, but I haven't really noticed anything bad while listening to music. I don't think there's much content in the final octave.
 
Fantastic project @Aijan . The vertical pattern control is amazing. I am surprised this project has not got the attention it deserves.

At what distance is the measurement taken?

I was also trying to understand your 2nd order crossover and expected a linear relation between the C-to-C and Xover frequency but could not establish a relation. I noted that :
1) D3/D6 is crossed at 18.9K and distance from edge to edge is 23cms (18cms ctc)
2) D2/D7 is crossed at 10.7K and distance from edge to edge is 35cms (30cms ctc)
3) D1/D8 is crossed at 6.1K and distance from edge to edge is (47 cms ctc)
Can you please elaborate how did you arrive at the crossover frequencies as related either edge to edge or ctc distance?
 
I am trying to make similar array using 12 of 4" inch coax drivers.
So to get a firm ground I started by simulating the OP's original 12 driver array using the same co-ordinates and low pass filters and got the below directivity. As expected it works well from 1Khz onwards.
1763025452028.png


To move ahead with my larger array, I also calculated the low pass filters vs wavelengths for the OPs 8 driver array and computed similarly for my 12 driver too as below
1763025495152.png


However, I expected good directivity performance between 333Hz (since its 3 times the OP's array's height) and 5Khz due to larger 10cm CTC than OP's 5cm.
However, the polars are not good, can anybody explain why?
1763025520030.png

Please ignore non standard component values, its just to get the simulation going.
Thanks and Warm Regards,
WonderfulAudio
 
Can you please elaborate how did you arrive at the crossover frequencies as related either edge to edge or ctc distance?
Sorry for not replying earlier. Unfortunately, I missed the email notification for your reply. I used VituixCAD simulations to come up with the crossover frequencies. The measurements were taken at 1 meter distance outside.

CTC distance is critical to minimize disruptive effects. You might want to try the CBT approach where they use increasing filter orders to induce a phase shift to make a straight line array behave as if it's slightly curved. That may improve the polar response.
 
Back
Top Bottom