28 



BELL SYSTEM TECHNICAL JOURNAL 



very different units, tlie 1X26 and the 1N25. Though direct comparison of 

 power handling ability is complicated by the fact that the burnout test 

 methods employed in the de^•elopment of the two codes were widely different, 

 it may be stated conservatively that while the 1X26 would be damaged after 

 absorbing something less than one watt peak pulse power, the 1X25 unit 

 will withstand 25 watts peak or more. The 1X26 unit is, however, capable 

 of satisfactory operation as a converter at a frequency of some 20 times that 

 of the 1X25. These two units have been made by essentially the same pro- 

 cedures, the difference in properties being principally due to modification of 

 alloy composition, heat treatment, and contact area. 



u 6.0 

 a. 7.0 



(j 8.0 



6.2 

 7.2 

 8.2 

 9.2 



notes: I. TEST FREQUENCY = 3000 MEGACYCLES 



2. NOISE RATIO IS THE RATIO OF THE AVAILABLE OUTPUT 

 NOISE POWER OF THE CRYSTAL RECTIFIER TO KTB 





IMPROVED PROCESS 



3.1 ^ 



■^ 6.3 



> 7.3 



O 8.3 



^ 9.3 



INITIAL PROCESS 



2.3 

 3.3 



0.1 0.2 0.4 0.6 1.0 2 4 6 8 10 20 40 60 100 200 



AVAILABLE PEAK PULSE POWER IN WATTS 



Fig. 16 — Comparison of Uie radio fre([uency power handling aljilit_\- of silicon crystal 

 rectifiers prepared by different processes. 



Prior to the process developments described above, in the interests of 

 simplifying the field supply problem one general purpose unit, the type 1X21 , 

 had been made available for field use. However, it became obvious that the 

 advantages of having but a single unit for field use could be retained only at a 

 sacrifice in either power handling ability or high-frequency conversion loss. 

 Since the higher power radar sets operated at the lower microwave 

 frequencies, it seemed quite logical to employ the new processes to improve 

 power handling ability at the lower microwave frequencies and to impro\e 

 the loss and noise at the higher frequencies. A recommendation accordingly 

 w^as made to the Services that different units be coded for operation at v^OOO 

 megacycles and at 10, ()()() megacycles. The decision in the matter was 



