278 



BELL SYSTEM TECHNICAL JOURNAL 



B. Negative Bias. — The curve of Fig. 3 shows the effect of negative 

 bias. As in the case of positive bias above, this curve is also an average 

 of three test curves. 



With this type of distortion the accuracy of reception is influenced 

 only slightly with increase in distortion over wide limits. There is a 

 gradual decrease in accuracy of reception up to about 65 per cent 

 distortion and a more rapid decrease beyond this value. 



100 



96 



=?92 



i;66 



84 



fiO 



10 20 30 40 50 60 70 80 



Per Cent Disfcorbion 



Fig. 3. 



It was thought that possibly the difference in the effect of positive 

 and negative bias was due to the adjustment of the sounders. If the 

 operators had adjusted the sounders to be biased slightly heavy for the 

 condition of zero bias over the circuit, then the sounders would have 

 failed sooner on a circuit with positive bias than on one with an equal 

 amount of negative bias. Some tests were, therefore, made for which 

 the sounders were adjusted so as to fail to operate properly for equal 

 values of large positive and negative bias in the circuit but the results 

 obtained were the same. It may be concluded, therefore, that the 

 effect is mainly of a psychological nature and is not due to a difference 

 in the adjustment of the apparatus. 



The noteworthy features in the results of tests with negative bias 

 are as follows: The accuracy of reception decreased gradually with in- 

 crease of distortion up to about 65 per cent, and fairly rapidly above 

 this value. The operators remarked upon the bias when it reached a 



