Computer-Aided Design of Loudspeaker Crossover Networks

Adams GJ, Roe SP. Computer-Aided Design of Loudspeaker Crossover Networks. J Audio Eng Soc. 1982;30(7/8):496-503. Available from: https://aes.org/publications/elibrary-page/?id=3830

@article{Adams1982_3830,
  author = {Adams, Glyn J. and Roe, Stephen P.},
  title = {{Computer-Aided Design of Loudspeaker Crossover Networks}},
  journal = {Journal of the Audio Engineering Society},
  volume = {30},
  number = {7/8},
  pages = {496--503},
  year = {1982},
  month = aug,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=3830}
}

TY  - JOUR
TI  - Computer-Aided Design of Loudspeaker Crossover Networks
AU  - Adams, Glyn J.
AU  - Roe, Stephen P.
T2  - Journal of the Audio Engineering Society
J2  - J. Audio Eng. Soc.
VL  - 30
IS  - 7/8
SP  - 496
EP  - 503
PY  - 1982
DA  - 1982/08/06
UR  - https://aes.org/publications/elibrary-page/?id=3830
PB  - Audio Engineering Society
LA  - en
AB  - A design technique is described whereby the component values of a passive loudspeaker crossover network are selected by a digital computer such that the value of some specified function is minimized. The functional chosen for the purpose of illustration represents the difference between some desired sound-pressure amplitude-frequency response and the calculated sound-pressure amplitude-frequency response of one component system of a multiple-driver loudspeaker system. Minimization of this functional is carried out using a numerical optimization technique implemented on a PDP-11 computer. To ensure that an accurate prediction of the frequency response of the system and the network combined is used in the optimization procedure, the calculation of the functional takes into account the measured sound-pressure-frequency response and the input terminal impedance of the component loudspeaker system. As a result of including these factors in the design procedure, the traditional requirement for trial-and-error adjustment of the values of the actual crossover components is shown to be eliminated. Because the functional is evaluated numerically at a number of discrete frequencies, the desired frequency response may be specified by any function of frequency. The technique is illustrated with several design examples.
ER  -