410 BELL SYSTE^f TECHNICAL JOURNAL 



Herschel. This structure passes low frequencies but attenuates strongly 

 frequencies for which the difference in path length of the two tubes is an 

 odd number of half wave-lengths. 



It is interesting to note that all of the fore-runners of the filter were of 

 the continuously distributed type which had their elements distributed 

 uniformly along the length of the device. All such devices have an infinite 

 number of pass bands, usually harmonically related. This is true also of the 

 dissipationless loaded line considered as a filter structure. It was only by 

 such abstractions as neglecting the weight of the loaded string that single 

 pass bands were obtainable. 



The low-pass electrical filter grew out of a network to simulate the opera- 

 tion of a long cable. By using series coils to simulate the loading and the 

 distributed inductance of the line, condensers to simulate the distributed 

 capacitance of the line, series resistances to simulate the ohmic resistance 

 of the line and the resistance of the coils, and shunt resistances to simulate 

 the leakance of the line, Campbell was able to obtain in a small space, a 

 device which had the same propagation characteristics up to the cut-off 

 frequency as a long section of loaded cable. Furthermore, by making the 

 resistance small, he was able to obtain a frequency range from zero fre- 

 quency up to a cut-ofif frequency fc with small attenuation, and a high at- 

 tenuation at higher frequencies, and thus obtained the first true low-pass 

 filter. He also put his filter to practical use for he says' 'T have made use 

 of these results by employing artificial loaded lines for cutting out harmonics 

 in generator circuits. The harmonics may all be cut down as far as desired 

 by the use of a sufficient number of sections, while the attenuation of the 

 fundamental can be reduced at pleasure by decreasing the resistance." 



Having developed the fundamental idea of a filter as a device for trans- 

 mitting without loss one frequency range and attenuating all other fre- 

 quencies, he went on to extend the idea to other types of filters in which 

 different frequency ranges were passed. The band-pass filter was already 

 realized in 1903 for Campbell says' "Combining condensers and inductances, 

 we may make a system which will not only cut out higher frequencies, but 

 also.all frequencies below a certain limit. " The high-pass and band elimina- 

 tion filters followed shortly after. With the invention of the electrical 

 wave filter, electrical network theory can be considered as well started. 



The science of electrical networks did not progress much farther 

 for a number of years. This was due primarily to the lack of application for 

 any of the structures developed. However, with the development of carrier 

 current transmission on telephone lines, the necessary stimulus was received. 

 Carrier current systems were an adaptation of radio communication systems 



" "On Loaded Lines in Telephonic Transmission," Philosophical Magazine, Ser. 6, 

 Vol. 5, pp. 313-330, IMarch 1903. 



