42 schunk: hysteresis of indicating instruments 



THE particular RESULTS OBTAINED 



Figure 2 presents the results of the cycHc calibration of a scale 



of the type described having the following general constants: 



Range of graduation: o to 20 pounds, corresponding to rotation 

 of the pointer through 2 revolutions = 4x. 



Number of joints having relative motion, 7 (including duplicates as 

 below; actual number of single journals and bearings 10). Turning 

 pairs : duplicated, 5 ; not duplicated, i . Higher pairs toothed gearing, 

 I. 



Displacement, for the purpose of the experiment, of the point of 

 application of the load from the line of centers of the link or stem sup- 

 porting the platform, 7.0 cm. 



Interval of graduation: 1/2 ounce = 1/32 pound = 1/640 or 0.0016 

 (approx.) of the nominal capacity. 



Interval of estimation of pointer position for the purposes of the 

 experiment: o.i ounce = 1/3200 or 0.0003 (approx.) of the nominal 

 capacity. 



After taking the readings reproduced in the figure, the drift of indica- 

 tion at full load was examined and found to amount to but 0.5 ounce 

 in nine and one-half hours. On removing the load the zero reading 

 exhibited a drift of 1.2 ounce to practically its initial value in i hour. 

 The amount of the drift occurring during the course of the series of 

 hysteresis observations recorded was probably negligible, and certainly 

 at no time greater than o.i to 0.2 ounce. The existence of drift in 

 directly measurable amount in this scale is believed to be the result 

 of an easily remediable structural defect, viz, weakness of a certain 

 portion of the inelastic sheet metal framework supporting the mechanism. 



The following are the particular results, many of which ap- 

 pear directly in the graph: 



I. The hysteresis loop, ^ or any portion of it, except at such a 



critical region as that designated BC in the figure, determined 



under the condition of slow^ aperiodic change of reading, was 



practically reproducible within a maximum deviation of one- to 



two-tenths of a graduation or at the largest separation of the 



loop where the relative precision is high, within a precision of 



about one or two parts in one hundred. It is to be noted that 



this small amount of dispersion of readings corresponds to a 



mutual displacement of repeated readings of but one-half the 



■* It is to be noted that the hysteresis obtained is for the major part the result of 

 frictional resistances and, concomitantly, of bearing clearances. (See Bureau of 

 Standards Scientific Paper No. 328.) It is known that only a relatively very small 

 portion of the hysteresis observed can be the result of elastic hysteresis of the re- 

 sisting spring of the scale. 



