A HIGH PRECISION STANDARD OF FREQUENCY 



497 



The manner in which the ring-shaped crystals are mounted in their 

 operating position is shown in Fig. 6. The hole is shaped so that 

 when the crystal hangs on a horizontal cylinder the point of contact 

 is at a theoretical node for mechanical vibration. There is evidence 

 of slight vibration where the central plane intersects the double 

 conical hole but it is small in comparison with that obtained at the 

 outer surface where a crystal is usually supported. The decrement 

 of the crystal when so mounted is considerably less than when it is 

 supported on one of its plane surfaces. 



In the mounting the crystal is spaced from the electrodes and is 

 kept approximately central by means of paper spacers on each side. 

 The crystal is free to move laterally in a narrow region but, since it is 



Fig. 6 — Section of crystal mounting showing point support. 



centrally located, the frequency is at a maximum value and hence a 

 slight motion of the crystal to either side has only a second order 

 effect on the frequency. 



The variation of frequency with total electrode spacing is appre- 

 ciable, but variations due to this factor are avoided by keeping the 

 electrodes accurately spaced by means of a ring of pyrex glass. The 

 temperature coefficient of pyrex is about one quarter of that of crystal 

 quartz perpendicular to the optic axis, so the variation in spacing that 

 is obtained is due almost entirely to the expansion of the crystal. 

 The effect on the frequency due to the differential thermal expansion 

 of the crystal and crystal holder is, however, less than one part in 

 10^ per degree C. and so it may be neglected. If it is desired to 

 eliminate this effect entirely a spacer should be used having the same 

 temperature coefficient of expansion as quartz perpendicular to the 



