A parametric array, in the field of acoustics, is a nonlinear transduction mechanism that generates narrow, nearly side lobe-free beams of low frequency sound, through the mixing and interaction of high frequency sound waves, effectively overcoming the diffraction limit (a kind of spatial 'uncertainty principle') associated with linear acoustics.[1] The main side lobe-free beam of low frequency sound is created as a result of nonlinear mixing of two high frequency sound beams at their difference frequency. Parametric arrays can be formed in water,[2] air,[3] and earth materials/rock.[4][5]
Priority for discovery and explanation of the parametric array owes to Peter J. Westervelt,[6] winner of the Lord Rayleigh Medal[7] (currently Professor Emeritus at Brown University), although important experimental work was contemporaneously underway in the former Soviet Union.[2]
According to Muir[8] and Albers,[9] the concept for the parametric array occurred to Dr. Westervelt while he was stationed at the London, England, branch office of the Office of Naval Research in 1951.
According to Albers,[9] he (Westervelt) there first observed an accidental generation of low frequency sound in air by Captain H.J. Round (British pioneer of the superheterodyne receiver) via the parametric array mechanism.
The phenomenon of the parametric array, seen first experimentally by Westervelt in the 1950s, was later explained theoretically in 1960, at a meeting of the Acoustical Society of America. A few years after this, a full paper[10] was published as an extension of Westervelt's classic work on the nonlinear Scattering of Sound by Sound.[11][12][13]
An alternate mathematical formalism using Fourier operator methods in wavenumber space, was also developed and generalized by Westervelt.[19] The solution method is formulated in Fourier (wavenumber) space in a representation related to the beam patterns of the primary fields generated by linear sources in the medium. This formalism has been applied not only to parametric arrays,[20] but also to other nonlinear acoustic effects, such as the absorption of sound by sound and to the equilibrium distribution of sound intensity spectra in cavities.[21]
Parametric receiving arrays can also be formed for directional reception.[29] In 2005, Elwood Norris won the $500,000 MIT-Lemelson Prize for his application of the parametric array to commercial high-fidelity loudspeakers.
To increase transmitting power and improve penetration, the signal must be amplified before being fed into the transducer. The Class-D power amplifier is widely used in the parametric loudspeaker due to its high efficiency and small size. However, Class-D power amplifiers are commonly used to amplify audio signals whose frequency is between 20 Hz and 20 kHz. When the PAA system is applied in underwater acoustic engineering and the signal frequency is higher, other amplifier configurations with broader bandwidth should be used. Svilainis suggested a power amplifier configuration and obtained 50 kHz to 3 MHz bandwidth [43]. The total harmonic distortion of 4% using 3 kΩ load and 400 Vp-p 1 MHz frequency signal was achieved, which indicated the ability to use such a power amplifier for high power and high frequency waveform excitation of transducers. Other researchers have developed power amplifiers for transmitting transducers by various techniques, such as digital predistortion and dynamic current biasing techniques to ensure high-voltage and low-distortion [44], power MOSFET linearizer scheme to improve the gain deviation characteristics of the power amplifier at higher input powers [45], etc.
In addition to the signal transmitting circuit, a complete PAA system in practice also requires a signal receiving circuit. As an example, Figure 8 shows the system configuration of a parametric array sub-bottom profiler. The system consists of a dry end and a wet end. The dry end includes a computer serving as a control and display center, and a battery serving as power supply. The wet end is composed of an embedded signal processor (DSP), a transmitting circuit, a primary frequency signal receiving circuit, a difference frequency signal receiving circuit, and a transducer array. The DSP generates the desired signal after receiving the instructions of the upper computer, and then the signal is sent out through the transmitting circuit, which comprises a digital-to-analog converter (DAC), some power amplifiers (PAs) and impedance matching networks (the number of which is in accordance with the transducer elements). In addition to the need for a high excitation voltage and satisfactory transducer elements, the successful transducer excitation also requires impedance matching between the transducer and the excitation source. In most cases, the equivalent impedance of piezoelectric transducers is capacitive. The parasitic input capacitance of the transducer clamps the PA output, and thus reduces the amplifier efficiency. If no electrical impedance matching network (EIMN) is adopted, it will not only seriously affect the power transfer efficiency, but also cause severe heating of the PA, resulting in permanent damage to components. There are many publications on EIMN design for piezoelectric transducers. The main design methods include analytical methods [46,47] and computer aided design (CAD) methods [48,49]. The analytical method is very complicated and requires the analytical form of the transducer, limiting its practice in engineering. The result of the CAD method is usually dependent on the choice of the matching network topology. A relatively simple and effective method is based on the Smith chart, assuming that the bandwidth has an inverse relationship with the quality factor Q [50]. Moreover, some researchers use optimization algorithms such as the genetic algorithm to search for optimal EIMN designs [51].
The phased parametric array is one of the most promising ways to achieve multi-beam scanning. However, during the beam steering process, the difference frequency wave of phased parametric array may have problems of broadening beam lobe and generating high intensity grating lobe and side-lobe. There are a few pioneer researches on the beamforming method for phased parametric array [52,69,80,81,82], but this technique is not matured enough to be of practice use. For a digital beamsteering system, the available steerable angle is often restricted by the sampling interval of the digital system. As the sampling frequency is not high enough, the smallest steering angle available is relatively large, making it difficult to distinguish signals in different directions. A number of fractional delay or frequency domain algorithms have been developed to improve the steering angle, but some of the algorithms requires high computational load and others introduce errors during the process. Gan proposed a digital beamsteerer for difference frequency in parametric array, which is able to steer to small angles without the need to increase the DSP board sampling frequency or implement fractional delay [80]. However, this method requires multiplications and additions for each sampling period, which to some extent increases computational complexity.
It's just a tiny little Purse Locker. Shouldn't have been too much trouble. Per previous posts on this topic, I first created the family, arrayed it in the X direction with two objects, and applied an array number parameter to it. I've even attached the file for your admiration and edification. So far so good, right?
Arrayed it in the Y direction. Piece #2 developed a mind of it's own and I had to sweat like the dickens to get it to sit where I want it. Applied an array number parameter - BOOM! Both pieces moved all over the place! Moved them (again), changed the array numbers to create the array. The X array worked perfectly. The Y array was a disaster. Please admire the screenshot and and instead of a box of tissue offer some sage advice....
I have a parametric array and I am trying to figure out the formula for the occasion where the array goes past a certain dimension, it would subtract one. My formula for the array is "Length/Spacing" where Length is 30' and spacing is 1'-4", which returns a Number of 23. This puts my last object outside of the 30'. I know that I can do "(Length/Spacing)-1", but that doesn't work in certain instances when the dimensions vary. I'm looking for a formula that would be along the lines of IF Spacing * Number is greater than Length, -1.
I just came to post that I finally found a formula that works. I created a parameter that returned a value of length x spacing, which I called "count". I made another parameter that I called "Overage" that returned the value of "Spacing * Count". My final formula for my array was "if(Overage > Length, Count - 1, Count).
What actually bothers you is the rounding of an array integer. For arrays, it is best to use a "rounddown" function. So, what is a rounddown? It rounds the number down. The round of 5.75 is 6, but the rounddown is 5.
@barthbradley , Good point, it does break when I enter a length of less than 3'.
@ToanDN , I started to look at curtain walls as a solution but didn't get that far into it. The end result here is actually framed walls, where there will be headers, trimmers, backing, etc. The array was actually for the studs. Is such a thing possible in a curtain wall?
@Ilic.Andrej, I will try the ROUNDOWN function. Thank you for that bit of information!
I started to look at curtain walls as a solution but didn't get that far into it. The end result here is actually framed walls, where there will be headers, trimmers, backing, etc. The array was actually for the studs. Is such a thing possible in a curtain wall?
I am testing out v19 to see if the new array options can be made to work for my situtaion. Specifically, I am working for a firm that makes chambers from panels that lock together. In my test, I drew a line to represent one wall, and then added a block to one end, to illustrate the panel. Then, I created an array of the block along the path, giving "measure" as the option. In this case, the panels are 36" wide.
aa06259810