268 Prof. J. A. Fleming on a Method of determining 



of the bar ; and this distance is very nearly equal to '56 of the 

 half-length of the bar from the centre or at '22 of the whole 

 length of the bar from one end. 



The following Table I. embodies all the results taken with 

 bundles of annealed transformer iron strip (Sankey's) of very 

 different lengths and sections, and Table II. similar observa- 

 tions on rods of steel and hard iron ; and it will be seen that 

 the effective induction in the case of the iron strip is always 

 found at a point in the bar or bundle which is very nearly 

 0*56 of the half-length of the bar from the centre. If 

 therefore the secondary coil is placed at that spot and the 

 secondary voltage then observed used to calculate the in- 

 duction density, the value so obtained is that which corre- 

 sponds to the true mean value of the varying hysteresis loss 

 per unit of volume all along the bar. In other words, if we 

 suppose the bar magnetized by an alternating current uni- 

 formly and with a maximum induction density everywhere 

 equal to that which it actually has under normal conditions 

 at 56 of the half-length of the bar from the centre ; then 

 the hysteresis loss in such a bar would be equal to that found 

 in the bar as it is. In all cases where induction density values 

 are given, these are to be understood as being the maximum 

 values during the cycle unless otherwise stated. The frequency 

 of alternations used was 80. 



We have therefore in this useful fact regarding the distri- 

 bution of induction in iron strips in a uniform magnetic field, 

 the foundation of a quick method of determining the hysteretic 

 constant of a sample of iron. 



If the iron sample, taken in strips, is placed as above in 

 the interior of the long magnetizing coil, and if a sensitive 

 wattmeter, suitable for use with alternating currents, is con- 

 nected therewith, we can measure the total power taken up 

 in the iron and in the copper of the coils when any known 

 alternating current is passing through the magnetizing coil. 

 If then the iron strips are withdrawn and another measurement 

 of the same kind made with the current passing through the 

 magnetizing coil adjusted to the same value as in the first 

 case, this last reading gives us the wattmeter-reading corre- 

 sponding to a known power taken up in the copper of the 

 magnetizing coils. Hence the first wattmeter-reading gives 

 us the iron loss, together with the copper loss, and the second 

 reading gives us the copper loss alone. The iron loss gives us 

 the true mean value of the hysteresis loss in the iron, assuming 

 the iron is so well laminated that eddy-current loss is prac- 

 tically absent. 



