1128 THE BELL SYSTEM TECHNICAL JOURNAL, SEPTEMBER 1956 



the waveguide wall with a series of disks which are closely spaced as shown 

 in Fig. 9 and the photograph of Fig. 10. The spacers serve a dual pur- 

 pose; to hold the disks in alignment and to provide loss for the spurious 

 modes. The circular disks provide the necessary continuity to support the 

 TEoi and TE02 modes and the gaps introduce high resistivity to the longi- 

 tudinal currents of the other modes. The spaced-disk filters, which were 

 arbitrarily designed to provide a 10 db loss to the TMn wave, were 1 %" 

 and 3 }/i" long for the Ke" ^i^d 14," waveguide sizes, respectively. In the 

 experiments to be described, a mode filter was inserted at each joint of 

 the line, at approximately 12-ft. intervals. , 



The measured attenuation data with mode filters at each joint of the 

 various fines are indicated by the dashed lines of Fig. 8. As shown the 

 effect of the mode filters is to reduce the TEoi loss variation by a factor 

 of at least two. 



The average attenuation is, however, generally somewhat higher than 

 for the unfiltered lines. This higher loss is partly due to spurious mode 

 power which is absorbed by the mode filter and is not reconverted to 

 TEoi power and to a slight degree to the increased TEoi loss introduced 

 by the mode filters. These results are shown in tabular form in Table II, 

 where the nomenclature is the same as in Table I. Because of the ex- 

 cellent performance of the 14" ^•^- hi^e (OFHC copper) by itself no meas- 

 urements with mode filters were performed on this line. 



CONCLUSIONS 



The measured data presented above indicate the feasibility of realizing 

 transmission losses as low as 0.5 db/100 ft. with the TEoi mode over dis- 

 tances up to several hundred feet. The transmission variations which 

 occur over the frequency band are a function of the circularity or tol- 

 erances of the waveguide. In a particular line the variations can be re- 

 duced considerably by adding mode filters along the line. It is reasonable 

 to expect that these variations can be reduced further by adding longer 

 mode filters at the joints or adding more mode filters at shorter intervals 

 along the line. Oxygen must be excluded from the line if the losses are to 

 be a minimum. 



ACKNOWLEDGMENT 



The author wishes to thank J. W. Bell and W. E. Whitacre for their 

 help in the measurements. 



This study was carried out at Holmdel and was sponsored in part by a 

 Joint Service Contract administered by the Office of Naval Research, 

 Contract Nonr-687(00). 



