78 



NATURE 



[March i8, 1920 



tion of the vertical ring, and O the lubber-line 

 support. 



By removing- the four screws marked n the gyro- 

 compass can be completely removed from the gymbal 

 rings. The instrument thus removed is shown in 

 Fig. 5. 



To explain the action of the oil bottles I have 

 introduced Figs. 6 and 7. 



Fig. 2 illustrates the simplest form of compass, in 

 which the wheel and case B are controlled by the 

 pendulous weight VV. When the case tilts, as shown, 

 W is moved to one side of the vertical support, and 

 the weight tries to bring the case again to the 

 horizontal. 



Suppose the wheel revolves in the direction of the 

 arrow a, the righting torque is in the direction of the 

 arrow b ; then the wheel and case will turn in azimuth 

 in the direction of the arrow e. Such a compass would 



other. At the middle of swing of the pendulum the 

 air-jet is at the middle of the air-box, and there is 

 no difference of air-pressure, and, therefore, no move- 

 ment of the oil; and when the swing is at the end 

 of its path, and not moving, the air-jet is at one side 

 of the air-box and producing the maximum move- 

 ment in the oil; it will therefore be seen that the 

 movement of the pendulum and that of the oil are 

 out of phase with each other. It is for this reason, 

 given good dynamic balance, that there is no quad- 

 rantal error whatsoever with this method of control. 



Fig. 7 illustrates the method of damping the 

 compass. Fixed to the same air-box K are the two 

 damping bottles C, C, smaller than E, E, but the 

 air here acts in the opposite direction to that in 

 Fig. 6. 



In one of these damping bottles is the adjustable 

 needle-valve, and this valve has a constricted passage, 



have a quadrantal error, because the weight W would 

 produce stresses in phase with the roll. 



Fig. 6 illustrates the method of control of the 

 Brown compass. When the case B is horizontal the 

 bottles E, E are half-full of oil, and the air-jet L is 

 blowing equally into the two halves of the air-box K ; 

 but when the case tilts, as shown, then the air- 

 pressure blows into one side of the box more than 

 into the other, and in such a direction as to force the 

 oil from the lower bottle into the one raised. There is, 

 therefore, a considerable righting torque indicated by 

 the weight of the oil W trying to restore the case 

 back again to the horizontal. 



When the pendulum swings under the action of the 



rolls of the shin, the air-jet L moves from one side 



to the other of the air-box in tune with the roll, 



blowing: the oil periodically from one bottle to the 



NO. 2629, VOL. 105] 



and thus the flow of oil from one bottle to the other 

 is suitably retarded. 



As regards the accuracy of the compass, I may 

 mention that one on board a flagship in the North 

 Sea during the war was observed with particular care, 

 especially during very heavy weather, and it was 

 reported that it was never more than i|° from the 

 true north position during the whole of the tests. - 



Trials on a commercial ship have demonstrated the 

 fact that the employment of a gvro-compass resulted 

 in the ship steaming every day more than 3 per cent, 

 p^reater mileage; in other words, one day's steaming 

 in thirtv would be saved, resulting in a proportionate 

 saving in coal and all other expenses. 



I come now to a most important application of the 

 fvro-compass, namelv, its employment as a gun 

 director. 



