r()TM'M:i) WAVK TIIKOUY A\D AVAVKCriDK APrMr'ATTOXS 



r()8 



measured forward and backward wave patterns in the roinid guide, for 

 excitation in one of the rectangular guides of the transducer. Only TEsi 

 of the six modes possible in the 2" pipe at 3.3 cm has a six-lobed pattern 

 of azimiithal distribution of radial electric field at the wall, and hence 

 the clean pattern with equalU^ spaced deep nulls indicates the pre.sence 

 of a rather pure TE31 mode. The six maxima of the forward wave were 

 eciual within ±0.15 db. The backward Avave had a peak electric field at 

 least 23 db down on the peak electric field of the forward wave. 



Using coupled transmission line technirjues and the familiar geometric 

 taper techniques, transducers were built for all of the six modes possible 

 in 2" diameter pipe at 3.3 cm for use in the circular electric \\ave research 

 program.^ These transducers were used to measure the forward wave and 

 backward wave output of the TlNIu transducers, as given is Table II. 

 In reality, imperfections in either one of the two transducers involved 

 in a measiu-ement could result in the recorded A^alues of discrimination. 

 For example, if the TMn transducer were perfect and the TEm output 

 transducer contained some T]\Iii . then the insertion loss measurement 

 involving the two transducers face to face would produce an indication 

 of mode impurity. Since we do not have independent information on the 

 mode ]:)urity of any one of the transducers at the level of the obser\'ed 

 wave im{)urities, we can only state that both transducers involved in a 

 discrimination measurement are probably at least as good as the number 

 tabulated. 



It should l)e noted that very high discriminations between TEoi and 

 TMu were achieved, despite the fact that this one discrimination de- 

 pends solely on the mode-.selective nature of the coupling orifice. Similar 

 discriminations can be employed effectively to augment the wave-inter- 



