Mon Jan 22, 2018 1:01 am
Before now, I haven't had a chance to get a good look at these speakers aside from the images that were already online, which didn't show any details of what was behind the grilles. I once met Ben Cooper of USSI, and asked him about these speakers, but he was tight-lipped on specifics (as he ought to be, being an engineer and protecting the 'secret sauce'). Looking closely at the newest renders of these systems, it looks like they are using cone drivers. The mystery was hidden behind their grille foam, but now that there are some images online without the grille foam, I can see that the acoustic elements are cone-type drivers. That is the secret to the overall package being so small; if these speakers employed horns, they would need to be larger, and all of the LRAD and Ultra Electronics systems employing horns are larger. For the little Ultra Electronics MA-Micro, which has a single driver facing in each of the 5 directions, the maximum SPL figure is 129 dB. What Ultra Electronics does not do is specify the distance at which the speaker was measured, so it can probably be assumed (at one's own risk) that the distance is actually the standard 1m reference distance used in loudspeaker design.
Looking up a high quality 6.5" midrange speaker using a neodymium magnet structure (high force compared to conventional ferrite magnetics), I found one (Ciare NDK6-1.5) that produces a peak of 108 dB (measured at 1m) at 4 kHz with 1 watt input; at the nominal power rating of 120W RMS it can theoretically produce 128.7 dB, which rounds up (for spec sheet purposes) to 129dB. Broadband, its sensitivity is lower - above 500 Hz, the average is 100dB. This would be an ideal driver for use in a mass notification system, because the response peak produced by resonant breakup of the cone is at the top of the voice range. Two of these drivers could produce 134 dB, each given 120W (240W total). Compact high frequency horn devices are available that could help these drivers above 4kHz, so take two of these in a sealed enclosure with a good HF horn and you can attain 134.7 dB at 1m. For our spec sheet, let's play marketing department and increase the number to 135 dB.
Where the siren starts pulling away is in distance performance. Sirens are not rated at 1m; they are rated at 100ft, or roughly 30m. A siren that produces 135 dB at 100ft (P-50) doesn't produce 135dB at 1m; an equivalent point source acoustic radiator would need to produce 164.6 dB at 1m to deliver 135 dB at 100ft. It is the need to produce such exceptionally high SPLs that requires a 50HP motor and a large blower fan inside a 6-foot horn.
So what can be expected from the speaker, compared to the siren? A speaker that can produce 135dB at 1 meter has only dropped off to 105.5 dB at 100'. If we say it is acceptable for the system to deliver coverage levels of 70dB at distance, as long as intelligibility is good, then the 135dB speaker system can throw out to 5,975 feet or 1-1/8 mile, neglecting atmospheric losses. Because the power handling of the speaker is a thermal phenomenon, we can even over-drive the speaker for brief transients, using delayed compression or true RMS limiting to give the speaker twice its rated power (or even more) for brief periods, such as the transients in speech, because it takes time for the speaker voice coil to heat up (though not much), so our speaker should be coupled to an amplifier that can supply 500W, with limiting provided by upstream processing. In that case, our speaker can actually produce brief transients at up to 138 dB (1m, 4kHz), improving intelligibility at long distances. Over the rest of the frequency band (and particularly at lower siren frequencies, where we are talking about 400 Hz instead of 4 kHz) the sensitivity of the speaker is 8 dB lower, and with a continuous tone at 400 Hz, the speaker can only throw a distance of 2350 feet, or about 0.45 miles, before the minimum threshold of 70dB is reached. If multiple speakers configured omnidirectionally are used, then the math changes slightly, because two speakers playing at equal levels will sum to a level that is 6 dB higher than one speaker playing alone. In this case, the maximum distance that two of these speakers may be spaced apart is 4700', or 0.9 miles. For voice notification on a college campus this is fine; for covering an entire town and incorporated area the number of devices starts to be excessive unless they are low in cost
The speaker we have been talking about so far is roughly on par with the Ultra Electronics MA-1; for the MA-2, distances can double. Even though the speakers are configured as a line array, at these long distances the speaker behaves as a point source and the inverse square law is in play, with the line array configuration only providing vertical pattern control.
Why do they sound so different? Using direct radiating cone drivers intended for audio applications, these systems probably produce distortion that is an order of magnitude lower than a horn-loaded system, which experiences resonances in the horn bell and has a narrower frequency response range, plus higher harmonic distortion caused by the compression loading of the drivers (throat overload); these are the price paid for extremely high output. They are also limited by not using horn loading; their effective radius is shorter compared to horn-loaded systems (which all electronic sirens are). Also, cone diaphragms made from paper are subject to UV degradation and mildew; the diaphragm of the driver chosen for this comparison has a cone treated with a weatherproofing PVA sealant, but it is not clear how long the driver will last under temperature, humidity, and UV exposure. I suppose that USSI/Ultra Electronics has already done the needed environmental testing to suitably qualify their chosen driver for outdoor installation, but paper materials are at a disadvantage for ultimate longevity; see the sirens that have been installed for 30 years or more on Hawaii for an example of what a phenolic-treated fabric driver diaphragm (typical of Atlas Sound siren drivers) can withstand. I suppose it's possible USSI could be using a thin Mylar diaphragm instead of paper.
Has Ultra Electronics discovered the technology that can replace electronic sirens as we know them? At this point I don't think so, but they have discovered a system configuration that works very, very well for its intended use: providing intelligible messages over a long distance. Were I designing a system using these speakers, I would install them near city centers, leaving the outlying areas to be covered by conventional electronic sirens.
Sorry this post is so long and is bringing back an older thread, and sorry it is such a brain-dump; I just happened to have an eureka moment about these systems.