THE L3 SYSTEM QUALITY CONTROL IN MANUFACTURE 985 



becomes sufficiently skilled in the terminating methods. Thus, in order 

 to obtain a product having normal distribution, close control of the 

 \\inding operation must be maintained. 



In addition to the above-mentioned single layer type coils, there is 

 another general family of coils wherein distribution controls are useful 

 during the initial period of development. The 1507 type, referred to as 

 L-R inductors, are precise low-Q single-layer inductors wound with re- 

 sistance \\dre on closely-controlled, low temperature-coefficient core 

 tubes. These inductors are used in the input feedback network and as 

 plate-feed inductors in the L3 amplifier, and are units which combine 

 resistance and inductance in a single product. The stability of these coils 

 is obtained by mnding on a ceramic core form, similar to that used in 

 the "splitting" coil, and using a low temperature-coefficient resistance 

 vdre in the winding. Simultaneous control of the resistance and induct- 

 ance of this coil is obtained by the calibration of the winding machine 

 setup to fit the resistivity of the ^vire used on the coil. Once the required 

 turns and pitch have been determined for a given spool of wire, experi- 

 ence has sho^^^l that the resistance control will remain valid until the 

 ^^'ire from the spool is exhausted. 



The problem of making an accurate electrical measurement of a low Q 

 inductor is a difficult one to resolve and it is here that an accurate control 

 of testing methods by means of statistical analysis was employed. When 

 an inductance (Lo) and a resistance (R) occur as series elements and the 

 combination is shunted by the residual capacitance (C) across the bridge 

 terminals, the resulting measured inductance is equal to that of the 

 original inductance minus the product of the capacitance and the square 

 of the resistance. 



Measured inductance = Lo — CR^ 



The above equation is generally true in the case of coils where the in- 

 ductive reactance is small compared to the resistance. If standard test 

 procedures were followed to make the measurement of inductance, using 

 the familiar Maxwell Bridges,'^ an error would be introduced due to 

 residual bridge and test jig capacitance. This would make the measured 

 value considerably different from the coil inductance and would be de- 

 pendent upon this residual capacitance and the coil resistance which 

 could not be controlled with any degree of uniformity. Parallel resonating 

 capacitance techniques are inadequate for the accuracies required, since 

 the shunt equivalent of a very low-Q inductance is equal to that in- 

 ductance divided by Q^. Determination of inductance by this method, 

 while better than the Maxwell Bridge method, is subject to errors due 



