628 BELL SYSTEM TECHNICAL JOURNAL 



levels which have been observed for each type. The No. 264B, with a 

 mean level of 42 db below 1 volt, is 20 db quieter than the No. 239A 

 which it was designed to replace, and is the quietest of the filament 

 type triodes. The next quietest tube of this structure is the No. 101 D, 

 in which the elements are supported from a rigid glass arbor and the 

 filament is quite heavy, requiring one ampere of heating current. The 

 No. 215A is almost identical with the No. 239A except for a firmer 

 supporting structure which results in a 5 db improvement. The most 

 microphonic of the types listed is the No. 231D, which has a very fine 

 wire filament whose diameter is fixed by the requirement that the heat- 

 ing current be 0.060 ampere. 



If the filament is the chief source of microphonic noise in filament 

 type tubes, then it is to be expected that tubes having indirectly heated 

 cathodes will be much less microphonic, inasmuch as the cathode is an 

 extremely rigid member. An examination of Table III shows that this 

 is indeed true. The No. 244A and No. 247A types, in which no 

 special precautions have been taken to obtain quietness, are about as 

 quiet as the No. 264B Tube. In the No. 262A Tube, therefore, it has 

 been possible to reduce the microphonic noise still further, to 49 db 

 below 1 volt, by cementing the elements into rigid supporting blocks 

 of ceramic material. This tube is also quiet in other respects, notably 

 in its freedom from AC hum picked up from the cathode heater circuit.'' 



In comparing tubes having widely different electrical characteristics, 

 it is not quite fair to compare their noise output levels alone, for given 

 two tubes having the same noise output, the tube having the higher 

 gain can be used with smaller signal inputs and have no greater noise 

 interference in the output. Accordingly, another column is given in 

 Table III listing the equivalent noise input level which would produce 

 the observed noise output if the tube itself were perfectly quiet. The 

 ratio of this value to the signal input level is directly related to the 

 degree of microphonic noise interference which is effective in the out- 

 put of the tube. It is computed by adding the voltage gain expressed 

 in db, of the tube in the measuring circuit, to the microphonic output 

 level obtained experimentally. The value of this criterion is illus- 

 trated in comparing the noise interference produced by multi-element 

 tubes and triodes. Multi-element tubes as a rule have higher noise 

 output levels than triodes as may be seen by comparing the Nos. 245A, 

 259A, 283A, and 285A screen-grid and pentode types with the Nos. 

 244A, 247A, and 262A triodes. When account is taken of the higher 

 voltage amplification of these former types, however, the noise inter- 



' "Analysis and Reduction of Output Disturbances Resulting from the Alternating 

 Current Operation of the Heaters of Indirectly Heated Cathode Triodes," J. O. 

 McNally, Proc. I.R.E., v. 20, pp. 1263-83, August, 1932. 



