376 SUBMAEINE CABLE LAYING AND EEPAIKING. 



the damping varies considerably, being very great on a low 

 resistance shunt, but with the universal shunt the damping is 

 constant throughoub all conditions of sensitiveness of the 

 galvanometer. 



Moving- coil instruments are, of course, unaffected by external 

 magnetic fields or disturbances, however severe, and this is a 

 most valuable feature of the Sullivan instrument. No mag- 

 netic screen, as in the ironclad marine instrument, is required. 

 In the Sullivan instrument the steadiness of the spot is un- 

 affected by changes in the ship's course, the zero keeping 

 absolutely constant while the ship is swinging to any point 

 of the compass. This is a very great advantage, especially 

 when testing the insulation of short lengths of cable on 

 board. Testing work generally can be carried on during a 

 passage at sea without having to wait as formerly for fine 

 weather to get a steady spot to work by. 



From tests made by the Author, one of these instruments at 

 42 in. focus and shunted by a 10,000 ohm universal shunt with 

 multiplying power of 4, gave 334 divisions deflection with a 

 standard cell of 1434 volts. With a multiplying power of 1, 

 and the same cell, the instrument gave 122*5 divisions through 

 1 megohm or 245 through 5 megohm. Under these conditions, 

 with the universal shunt, the spot stopped at the deflection or 

 zero with no overs wiog. Tried with deflection off" scale, the return 

 to zero was without any overswing whatever. When the shunt 

 was disconnected altogether from the instrument the deflection 

 with the same cell was 144 divisions with 10 divisions over- 

 swing once oifly. To test for " creep," a deflection of 362 

 divisions (one cell through 0*4 megohm no shunt) was kept on 

 for five minutes, at the end of which the spot had crept 

 one division only, and the return to zero was only half a division 

 out. The proportionality of deflections with shunts of different 

 multiplying powers was exact, allowing for scale correction. 



Tested for signalling with the brush damper applied with one 

 cell through 3 megohm and universal shunt of multiplying 

 power = 1, the signals were quite sharp and easily readable at 

 considerable speed. The brush reduces the signals to a con- 

 venient size for reading, and does away with all wandering of 

 zeroc The effect produced on the signals is like a sharp blow 

 instead of a light blow followed by a push. 



