MEASUREMENT OF TELEGRAPH TRANSMISSION 173 



transmission testing consists in the addition of a small crank which 

 extends through the cover and which is coupled to the range finder. 

 The crank has a detent which assists in making settings to the nearest 

 per cent, the scale being arranged to indicate directly the distortion 

 in percentage of a unit selecting pulse. This crank and scale arrange- 

 ment increases the convenience of measurement considerably. 



It is of considerable importance to remove as far as practicable the 

 internal distortion ^ of teletypewriters used for measuring purposes. 

 These internal distortions can be identified as bias, characteristic and 

 fortuitous effects. These effects reduce the margin from the theoretical 

 limit of ±50 per cent to about ±40 per cent for the usual machines 

 operating at 60 words per minute. 



The presence of internal distortion can be readily determined by 

 using signals biased by various amounts and noting the effect on the 

 orientation range. With a machine satisfactory for measuring pur- 

 poses, the results obtained would be as indicated on Fig. 21a. Here the 

 range is from 10 to 90 per cent for reception of perfect signals. Mark- 

 ing bias reduces the upper range in direct proportion and spacing bias 

 likewise reduces the lower range in direct proportion ; thus the machine 

 would be satisfactory for measuring purposes. 



If the machine had internal marking bias, the orientation parallelo- 

 gram would be as shown in Fig. 2\h. Here the range with perfect 

 signals is from 10 to 80 and marking bias reduces the range in propor- 

 tion to the bias but spacing bias first increases the range until the 

 internal bias is compensated and then decreases the range as the 

 impressed bias is further increased. With internal spacing bias the 

 parallelogram would be shifted to the other side of the zero line by the 

 amount of the bias. It is obvious therefore that biased machines do 

 not give accurate results in measuring. 



Characteristic and fortuitous effects may also be present in tele- 

 typewriters to such an extent as to make the machines unsatisfactory 

 for testing purposes. It will be appreciated that internal characteristic 

 distortion changes from signal to signal and when receiving miscel- 

 laneous teletypewriter characters, it has much the same effect as 

 fortuitous distortion and the upper and the lower margin limits would 

 be reduced about equally as shown in Fig. 21c. If the machine 

 distortions do not have the same effect on mark-to-space and space- 

 to-mark transitions, a skewing effect is produced in the orientation 

 parallelograms. Fig, 2\d shows skew due to the fortuitous effect 

 of the mark-to-space transitions, being greater than the fortuitous 

 effect of the space-to-mark transitions. Teletypewriters showing 

 such skew effects do not give margin reductions proportional to the 



