294 
The N.Z. Journal of Science and Technology. 
[Jan. 
'£Zf 
CO 
portable instrument is described for obtaining this factor, and the instru¬ 
ment is now manufactured by the Leeds and Northrup Company. It will 
readily be seen that the same devices and arrangements which reduce 
induced voltage will also reduce telephone interference— e.g., transposition 
of the power circuit. In addition, the value of a pure sine wave will be 
evident; and, apart from the design of the generators, the two factors 
which produce avoidable harmonics are—(1) Suitable arrangement of trans¬ 
former connections so as to suppress the third harmonic and its odd 
multiples; (2) avoidance of too large a magnetic flux density in the 
transformers, as this would produce a flat wave with consequent harmonics 
A simple qualitative method of estimating the relative interference 
caused by different harmonics is shown in fig. 7. A telephone-receiver t 
is connected either directly to the line (the interference in which is to 
be investigated) or through a capacity c of about ^ microfarad to a 
lighting circuit derived from the power circuit. t is held over a tube 
resonator, the length of which can be adjusted to 
resonate to the different harmonics. The relative 
intensity of sounds produced by resonance to a par¬ 
ticular harmonic gives a qualitative measure of its 
interference effect. The frequency can be calculated 
V 
from the formula n = - . /7 , ——, where n is the fre- 
4A (l -J- -3 a) 
quency and V the velocity of sound = 330 -j- *6£° 
(centigrade) metres per second. Care must be taken 
that the resonance tube is giving its fundamental note. 
This can be confirmed by reference to a musical instru¬ 
ment. 
Electro-magnetic Effects. 
Currents in a communication circuit due to electro¬ 
magnetic induction from a power circuit are relatively 
less important as they are proportional to the current 
in the power line and not to voltage. Moreover, in the 
case of twin telegraph or telephone circuits the induced Fig. 7. 
E.M.F. on one wire will tend to balance that in the 
other, especially if the circuit is transposed. The interference becomes 
serious only in the case of an earthed telegraph circuit, and in this case it is 
induction due to the fundamental frequency which is more likely to affect 
the instruments owing to the inertia of the moving parts. 
As in the case of electrostatic induction, the resultant effect is equal 
to the vector sum of the effects of the three lines, and in the hypothetical 
case of balanced currents equidistant from the communication circuit will 
add up to zero. Calculations of the effects in practice have been made by 
Parry (this Journal , vol. 2, p. 308, 1919). 
With unbalanced circuits due to unsymmetrical loads or unequal leakages 
to earth from the lines there will be a residual current which will be 
relatively much more important as regards interference. The effect of 
this residual current, as in the case of electrostatic induction from residual 
voltage, will not be eliminated by simple transposition of the power circuit 
without co-ordinated transposition of the communication circuit. It is 
therefore necessary to keep the current in the various lines at the same 
value and the same phase relationship (120°). 
Question of Transposition. 
We have so far attempted to give only a few of the underlying principles 
governing interference. It will be obvious that the main essentials are 
