ESTIMATION AND CONTROL OF OPKK A TK TIME OF UKLAYS 151 



shows the successive steps taken in detei'iiiiiiiiiji; the iiumerical vahie for 

 a specific case. 



The dc circuit resistance is detei'mincd by the known circuit voltage 

 and specified power. Entering tiu^ chart on the left ordinate at this re- 

 sistance and proceeding horizonttdly to the right, intersections with 

 increasing number of turns lines, sloping downward to the right, are 

 found, and the appropriate one is chosen. Dropping vertically to the 

 abscissa, the winding time constant tc = LiGr is determined. 



To this is next numerically added the known effective core eddy cur- 

 rent and a sleeve (if any) time constant by returning vertically to an 

 intersection with the appropriate core indicated as "no sleeve," or core 

 plus slecN'e, curve. Proceeding horizontally to the right hand ordinate 

 scale from this intersection, the time const-int multiplier of the In term 

 in equation (4) is determined. 



The multiplication of these two factors is accomplished by proceeding 

 to the left along this same horizontal line, to an intersection with the 

 proper q line, sloping downward to the left. Vertically below this last 

 intersection is the operate time. 



Initially, tentative q lines are drawn, omitting the motion time cor- 

 rection. These lines are straight with a positive 45° slope. Then with an 

 actual relay whose just operate current has been measured, operate 

 time measurements are made, keeping the final current, and hence q, 

 fixed at se\'eral values in tvn-n. This is done by adjusting an external 

 resistance and battery voltage over a wide range, effectively changing 

 X~/R. These measured data are plotted, following the same steps through 

 the nomogram, except the last intersection is with the measured time 

 vertical, rather than the known q. This provides several empirically 

 determined q lines. These are used as templates, to progressively alter 

 the shapes of the tentative straight q lines drawn earlier and shift them 

 to the right. This adjustment then introduces the motion time correction. 



It has been found empirically that for large operating times, the motion 

 time correction factor has a vsdue of about 0.1 There is no definite 

 di\'ision between mass and load controlled operation, but as the total 

 time decreases, travel time becomes more important. The correction 

 factor increases to a value of about 0.5, in the transition range. For 

 completely mass controlled operation, there is no q effect, so the q lines 

 must all eventually con\'erge. 



As stipulated above, this chart applies to the current, ftux and pull 

 range w^here the first approximation magnetic constants of the structure 

 are applicable. For this reason, the tests for the motion time corrections 

 are made under conditions meeting these restrictions. 



These curves should be corrected for magnetic saturation if there is a 



