MEASUREMENT OF TELEGRAPH TRANSMISSION 157 



arm (resistances 3 and 7) of a bridge type of circuit containing recording 

 meter M. The two arms of the bridge are balanced and the meter, 

 being bridged across them, receives positive and negative current 

 pulses of equal magnitude in response to the armature movements. 

 If these pulses are of equal duration, as for telegraph dots or reversals 

 having zero bias, the average meter current will be zero. Biased 

 reversals will cause the meter current to average at other than zero 

 by an amount directly proportional to the percentage bias. A center- 

 zero recording meter is used and this provides a running record of the 

 variation in bias. A damping condenser C is used to reduce the 

 width of the trace and the amount of unsteadiness of the indication 

 due to fortuitous effects. 



The alarm circuit contains a galvanometer-relay bridged across equal 

 resistances 4 and 7. The needle of the galvanometer-relay moves to 

 one contact or the other when excessive values of bias are experienced, 

 the sensitivity being adjusted for response to different values of bias 

 by means of adjustable resistance 2. The response is made somewhat 

 sluggish to avoid alarms being given for interruptions of short duration 

 which are not of interest in connection with an investigation of bias 

 stability. 



A sample chart obtained by means of one of these sets is shown in 

 Fig. 9. This chart shows slow variations in bias in the upper part 

 and in the lower part the change in the indication due to dropping 

 out or adding a single dot and momentary failures. Such charts do 

 not, of course, show the characteristic distortion, since this is not 

 present in the case of unbiased reversals. 



These sets are now generally used in the field in routine checks on 

 telegraph circuits and in special checks on circuits which have de- 

 veloped faults in service. Usually these checks are made with the 

 idea of locating the cause of hits or swings which it is difficult to 

 locate otherwise ; in some cases sets are used simultaneously at several 

 repeater points to sectionalize trouble. Charts are run for long 

 periods, sometimes for several weeks in such tests. The stability test 

 sets are also used to obtain data for transmission ratings of circuits, 

 in which case it is desired to know the extent of the bias variations 

 over long periods and the number and frequency of occurrence of hits. 



Figure 10 shows a view of the portable arrangement of the set 

 including the recording meter. The alarm buzzer and the receiving 

 relay are located on the panel along with a row of keys for adjusting 

 the alarm for operation on given values of bias. Jacks are provided 

 on the side of the box for connection to circuits, batteries and the 



