the Winding of Voltmeters. 115 



1. Shielding the instrument by putting it in an iron case 

 with very thick walls. 



2. Using a very powerful controlling field. 



3. Using a motion of translation, and not of rotation, for 

 the body in the instrument that is deflected by the current. 



The first method to be successful requires that the iron 

 case should be heavy and cumbersome, and makes it unsuitable 

 for portable instruments. 



A very powerful controlling field necessitates a very 

 powerful deflecting force, and so either makes the instru- 

 ment unsensitive or necessitates a considerable expenditure 

 of electric energy in the instrument. The third method, 

 consisting in giving the deflected body a motion of translation, 

 is the best, since, no matter how strong the disturbing 

 magnetic field may be, provided it is uniform throughout 

 the small space traversed by the deflected body on its motion, 

 no increase nor diminution of the deflection will be caused 

 by the outside magnetic field. For this reason we have 

 adopted this third method in our Magnifying Spring-instru- 

 ments, and experiment shows that the shielding produced in 

 this way is even better than in our older form of instruments 

 provided with a powerful permanent magnet. 



X. As to the third error, arising from certain types of volt- 

 meters altering in sensibility so that their constants vary 

 from time to time. This error arises from two causes. In 

 some voltmeters, such as those in which a controlling per- 

 manent magnet is employed, there is an error which gradually 

 comes on from the demagnetization of this magnet, an error 

 which is greatly increased by a mistaken careful regular em- 

 ployment of a keeper for the magnet when the instrument is 

 not in use. This error can be avoided by dispensing with a 

 permanent magnet, and using a spring or a weight, or an 

 electromagnet which is magnetized by the current to be 

 measured. A second time-error arises from the employment 

 of iron in the neighbourhood, and has the effect of making 

 the indications of the instrument depend not merely on the 

 current passing at the time of measurement, but also on the 

 currents that have lately passed through the instrument. 

 This error is especially marked in those instruments in which 

 the controlling force, or the force tending to bring the needle 

 back to zero in opposition to the deflecting force, is produced 

 by the attraction of an electromagnet containing a core of 

 soft iron wire, and which it is assumed is magnetized to 

 saturation by the current which it is our object to measure. 

 With small currents, however, which produce deflections in 



12 



