308 BELL SYSTEM TECHNICAL JOURNAL 



the same width of radiated beam and resolving power. To meet the needs 

 for aircraft and submarine radar and other appHcations in which antenna 

 size must be small, shorter wavelength magnetrons were essential. The 

 development of magnetrons of frequency near 10,000 mc/s (3 cm. wave- 

 length) was part of the expanding program of magnetron research and de- 

 velopment which grew out of the earliest experiments. The work was 

 concentrated at 3.2 cm., roughly a factor of three below the initial 10 cm. 

 work. From the earliest 3 cm. magnetron developments was evolved an 

 efficient 3.2 cm. magnetron, the 725A, requiring roughly the same driving 

 conditions as were used in the first 10 cm. magnetron applications. 



An operating 3 cm. magnetron was built in the summer of 1941. The 

 crude techniques then in use in the 10 cm. wavelength range were adapted 

 to the 3 cm. range, and 5 kw. peak power from this magnetron was measured 

 in a coaxial water load. This design involved an unstrapped resonator 

 system having eighteen quarter wavelength slots. The choice of so many 

 resonators was made in order that large cathode and anode dimensions could 

 be used, later shown to be unreasonable for the voltage range employed. 

 In addition the design suffered from a confusion of many modes of the 

 resonator system and from an inadequate output circuit. 



A later design, similar in some respects to one upon which work was done 

 at the M. I. T. Radiation Laboratory, made use of a resonator system having 

 twelve slots. This design operated in a mode other than the tt mode, prob- 

 ably that for which n = 3 or n = 4. The M. I. T. version, the 2J21, was 

 the first 3 cm. magnetron to be manufactured in quantity. Its power output 

 was about 15 kw., generated at an efficiency of from 12 to 15 per cent. 

 Many variations of this design were made without major success in an 

 attempt to obtain efficient and reliable operation. 



Prior to the introduction of straps, attempts were made to adapt 10 cm. 

 magnetron designs to 3 cm. While none came up to the 10 cm. operating 

 efficiencies, one such, having an unstrapped resonator system of eight hole 

 and slot resonators, was used extensively in systems experimental work in 

 our Laboratories. 



By this time it was realized that some fundamental reason existed for the 

 failure of these 3 cm. magnetrons to reach efficiencies comparable with those 

 obtained at 10 cm. The resonator systems were unstrapped at 3 cm., and 

 they operated in a mode other than the tt mode. However, this did not 

 fully explain their failure. Attempts were made to operate strapped 3 cm. 

 magnetrons in the tt mode, but they still did not result in the desired or 

 expected improvements. The trouble lay not only in the mode frequency 

 distribution of the resonator system but also in the size of the interaction 

 space. Prior to this time, 3 cm. magnetrons were made with larger inter- 



