ELECTRICITY METERS 



517 



subdivision is done mechanically. The 20 per cent load mark is 

 the lowest one on the scale. 



The readings at the lower end of the scale are considerably 

 influenced by temperature. The effect of temperature becomes 

 smaller as the readings increase. This is illustrated by the fol- 

 lowing tests: Two indicators, one of 5, the other of 10 amp. 

 capacity, were used. They were placed in a suitable chamber in 

 which were heating coils and a fan to circulate the air. The tem- 

 perature being originally at 68F. was raised to 104F., no cur- 

 rent being sent through the indicators. The rise of liquid in the 

 index tube was for the 5-amp. indicator 0.38 in., for the 10-amp. 

 indicator 0.50 in. 



The temperature was maintained constant for 4 hr. at 104F., 

 the fan being kept running; at the end of this time the 20 per 

 cent, load test was begun ; the other tests followed as usual, and 



10 20 30 40 50 60 70 80 90 100 

 Per Gent of Pull Load 



FIG. 304. Illustrating the effect of temperature on two Wright demand 



indicators. 



the results are shown in Fig. 304. It will be seen that the low 

 readings are very considerably affected, that the percentage error 

 decreases with an increase of load, and that each instrument has 

 its own characteristic behavior. 



The leads to the indicator influence its action by conducting 

 heat away from the bulb; the error so introduced will depend on 

 the size of the leads and the difference in temperature between 

 the heater strip and the outside air. 



In a two- wire service one indicator is installed on the customer's 

 side of the watt-hour meter. According to American practice it 

 is customary in a three- wire service to install two indicators, one 



