556 



BELL SYSTEM TECHNICAL JOURNAL 



Sources of Error 



There are certain sources of error inherent in all electrodes which 

 affect both the reproducibility and the accuracy of the results ob- 

 tained. First, we have the question of contact between the electrode 

 and the sample. If the electrode does not make perfect contact 

 over its entire area with the sample, the result obtained is not a true 

 value for the material of the sample, but is a resultant, depending 

 upon the amount and nature of the material filling the gap. Air 

 spaces between the electrode and the sample have a marked effect 

 on the apparent dielectric constant. An air-gap. 001 in. thick in 

 series with a sample having a dielectric constant of 5 will have the 

 same effect on the capacitance as increasing the thickness of the 

 sample by .005 in. If the actual thickness of the sample is .05 in. this 

 results in an error of nearly 10% in the value of dielectric constant. 

 The power factor will also be reduced and the loss factor or product 

 of power factor and dielectric constant will be reduced by the factor 



( — y, or about 17%. Thus it is evident that the elimination of all 

 \55/ 



gaps between the electrode and the sample is one of the first require- 

 ments both for reproducibility and accuracy. 



A second effect inherent in all electrodes which is a source of error 

 unless properly allowed for, is the so-called fringing of the electro- 

 static flux, that is, the lines of force tend to spread out and include 

 an area of the sample greater than that of the electrode. This is 



Fig. 1 



illustrated in Fig. 1. So far as the flux which is confined entirely 

 to the material of the sample is concerned, this produces an error 

 in the dielectric constant, which involves a determination of the 

 effective area of the sample, but not in the power factor which is 

 independent of the area. However, there are some lines of force 



