6 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1952 



of resonance that the insertion loss of such a device would be prohibitively 

 large. 



McMillan in his original article, showed that a gyrator could be 

 realized by means of mechanically coupled piezo-electric and electro- 

 magnetic transducers. Later, McMillan pointed out that a gyrator 

 could be realized by means of the Hall effect in a square plate of bismuth, 

 as was also predicted by Casimir. Another similar possibility would be 

 an electrical-electrical coupling through a gyroscopic link. A gyrator has 

 been built by W. P. Mason of these Laboratories which makes use 

 of the Hall effect in a crystal of germanium. This gyrator showed an 

 insertion loss somewhat higher than the theoretical loss of 12.3 db. 

 R. 0. Grisdale of these Laboratories suggested that these losses could 

 be greatly reduced if the same Hall effect principle were applied to a 

 vacuum tube which contained four electrodes which could both emit 

 and collect electrons. This device is no longer passive, but such a struc- 

 ture has been built and showed an insertion loss of about 7 db, only 

 slightly higher than the theoretical loss which would be expected from 

 this geometry. 



In view of the substantial losses found to exist in the earlier forms of 

 gyrator discussed above, a study of other "anti-reciprocal" phenomena 

 which might lead to the realization of a relatively low loss gyrator was 

 undertaken. 



It has long been known that the Faraday rotation of the plane of 

 polarization in optics is anti-reciprocal. In order to observe the Faraday 

 rotation, polarized electromagnetic waves must be transmitted through 

 a transparent isotropic medium parallel to the direction of the lines of 

 force of a magnetic field. The effect is usually produced by placing the 

 material along the axis of a solenoid. The rotation is "positive" if it is 

 in the direction of the positive electric current which produces the field 

 and "negative" if in the opposite direction. All optically transparent 

 substances show the Faraday rotation. 



Its anti-reciprocal property distinguishes the Faraday effect from 

 optical rotations caused by birefringent crystals, or by the Cotton- 

 Mouton effect, which are reciprocal. That is, if a plane polarized light- 

 wave is incident upon a birefringent 'crystal in such a manner that the 

 plane of polarization is rotated through an angle 6 in passing through 

 the crystal, then this rotation will be cancelled if the wave is reflected 

 back through the crystal to its source. In the Faraday rotation, however, 

 the angle of rotation is doubled if the wave is reflected back along its 

 path. Hence, if the length of path through the "active" material is 

 adjusted so as to give a 90° original rotation, the beam on being reflected 



