Tweet
Originally posted by Rory Buszka
View Post
There were some requests for additional information about my DIY "LRAD" portable acoustic hailing demonstration, and the speaker components used for the "acoustic emitter" - the speaker head.
The drivers used were PRV Audio D4400Ph-Nd phenolic-diaphram neodymium drivers, with a combined weight of 35 lbs of driver in the speaker, plus about 10 more pounds of MDF and hardware. They are mounted to horns from "Blast King", a car audio company with a cheeky name, but they were the right size and the right flare rate, and they had a very solid thick wall construction. The D4400Ph-Nd is about 6" in outer diameter, and they have a 4" diaphragm with a 4" voice coil, so they are the most monstrous midrange-oriented compression driver I could use at a reasonable price to keep the project feasible. (BMS has a dual-diaphragm midrange compression driver but that would be unreasonably expensive.) The emitter was positioned a little less than 6' off the ground using a pro-audio tripod speaker stand.
Peak sensitivity of the driver is 113dB 2.83V/1m, and the four drivers operate as a coherent 'phased array' summing to provide a voltage sensitivity of 125 dB @ 2.83V/1m, with the impedance dropping to 2 ohms (all four drivers are in parallel). This is no problem for the amp used, which is an auto sound amp with an internal step-up power supply, designed to draw power from a 12V battery. The amp I used for the demonstration was a SounDigital Nano 250.2 supplying 125W into the 2-ohm load, but the plan for future demos is to use a Taramps HD3000 2-ohm variant to supply up to 3kW peak into 2 ohms.
Frequency response is from 500 Hz at -10dB to 7.5kHz at -10dB (full power operating range 400-8,000 Hz should be no problem), and peak SPL is at 4.5kHz, with smooth-ish response between 1kHz and 5kHz (no big peaks).
With the amp I used, peak SPL at 4.5kHz/1m should have been around 145 dB, and SPL at the 500' distance of the listening position should have peaked around 100 dB, but for various reasons (including atmospheric loss, windage, ground reflections, power compression, throat distortion, or the lower crest factor of speech compared to tone signals) I suspect I wasn't achieving that level, at least not continuously. There is more testing to be done. With the bigger amp, it should be possible to reach continuous on-axis levels of 154 dB at 1m, and peaks of 158 dB; that's pretty lofty so I'll need to figure out a way to validate that. The guidance from FEMA for designing outdoor warning siren systems uses a -10dB per distance doubled model instead of -6dB for a typical point source, so they are factoring in more losses. Even so, all reports during the demo confirmed there was no lack of volume.
Thanks for indulging me with the chance to demonstrate this project publicly. There's nothing really patentable here; the design approach is to implement a modular construction with a phased array of 'cells' that provide good efficiency in converting electrical power to acoustic power. They can then be configured in any manner needed to optimize dispersion and suit the application, such as a panel array (imagine 2x4 or 3x3 units in a grid), a circular array of 12 drivers, a hexagonally-packed array on a flat surface, or a ring array (annular array) where very narrow dispersion is needed from a minimum number of sources.
JR's wife took a video: https://www.youtube.com/watch?v=OhusYx4Dv3Q
The drivers used were PRV Audio D4400Ph-Nd phenolic-diaphram neodymium drivers, with a combined weight of 35 lbs of driver in the speaker, plus about 10 more pounds of MDF and hardware. They are mounted to horns from "Blast King", a car audio company with a cheeky name, but they were the right size and the right flare rate, and they had a very solid thick wall construction. The D4400Ph-Nd is about 6" in outer diameter, and they have a 4" diaphragm with a 4" voice coil, so they are the most monstrous midrange-oriented compression driver I could use at a reasonable price to keep the project feasible. (BMS has a dual-diaphragm midrange compression driver but that would be unreasonably expensive.) The emitter was positioned a little less than 6' off the ground using a pro-audio tripod speaker stand.
Peak sensitivity of the driver is 113dB 2.83V/1m, and the four drivers operate as a coherent 'phased array' summing to provide a voltage sensitivity of 125 dB @ 2.83V/1m, with the impedance dropping to 2 ohms (all four drivers are in parallel). This is no problem for the amp used, which is an auto sound amp with an internal step-up power supply, designed to draw power from a 12V battery. The amp I used for the demonstration was a SounDigital Nano 250.2 supplying 125W into the 2-ohm load, but the plan for future demos is to use a Taramps HD3000 2-ohm variant to supply up to 3kW peak into 2 ohms.
Frequency response is from 500 Hz at -10dB to 7.5kHz at -10dB (full power operating range 400-8,000 Hz should be no problem), and peak SPL is at 4.5kHz, with smooth-ish response between 1kHz and 5kHz (no big peaks).
With the amp I used, peak SPL at 4.5kHz/1m should have been around 145 dB, and SPL at the 500' distance of the listening position should have peaked around 100 dB, but for various reasons (including atmospheric loss, windage, ground reflections, power compression, throat distortion, or the lower crest factor of speech compared to tone signals) I suspect I wasn't achieving that level, at least not continuously. There is more testing to be done. With the bigger amp, it should be possible to reach continuous on-axis levels of 154 dB at 1m, and peaks of 158 dB; that's pretty lofty so I'll need to figure out a way to validate that. The guidance from FEMA for designing outdoor warning siren systems uses a -10dB per distance doubled model instead of -6dB for a typical point source, so they are factoring in more losses. Even so, all reports during the demo confirmed there was no lack of volume.
Thanks for indulging me with the chance to demonstrate this project publicly. There's nothing really patentable here; the design approach is to implement a modular construction with a phased array of 'cells' that provide good efficiency in converting electrical power to acoustic power. They can then be configured in any manner needed to optimize dispersion and suit the application, such as a panel array (imagine 2x4 or 3x3 units in a grid), a circular array of 12 drivers, a hexagonally-packed array on a flat surface, or a ring array (annular array) where very narrow dispersion is needed from a minimum number of sources.
JR's wife took a video: https://www.youtube.com/watch?v=OhusYx4Dv3Q
Comment