tion into ways of improving reliability and 

 efficiency of underwater lights. Table 10.3 

 was taken from Waltz's report and it summa- 

 rizes the characteristics and pro's and con's 

 of the commercial light types discussed 

 above as well as others. 



On the whole, the NURDC study confirmed 

 the results of Strickland and Hittleman, add- 

 ing that there was a lack of demonstrated 

 reliability in all existing light systems. Light 

 failures occurred both during operations and 

 during design assurance testing. The pri- 

 mary type of failure during operation was 

 stated to be leakage of water into the glass 

 envelope housing which usually caused a 

 short circuit in the power supply lines. 



Analyzing the mercury vapor light failures 

 on the DSRV, Waltz noted that leakage was 

 not the problem, but, instead, the fluctuating 

 power supply from the DSRVs main batter- 



ies was a chief culprit. The mercury lights 

 are designed for AC operation, hence, the 

 power must be converted from DC to AC and 

 then current limited, this is accomplished 

 within a ballast unit. The ballast units did 

 not regulate the varying voltage input (90 to 

 140 VDC) adequately and resulted in the 

 light being driven at a greater power level 

 than its nominal rating at high inputs (which 

 reduced its lifetime) or driven at a lower 

 than rated voltage which decreased its lumi- 

 nous efficiency. The prime DSRV ballast unit 

 deficiency was in the electrical power conver- 

 sion efficiency which was as low as 55 per- 

 cent. This manifested itself in the form of 

 heat which caused failure of several units. 

 Filling the ballast units with a dielectric oil 

 to improve cooling efficiency and replacing 

 some electric components solved the problem 

 somewhat. 



Another serious deficiency with the mer- 



TABLE 10.3 SUMMARY OF LIGHT SOURCE CHARACTERISTICS [FROM REF. (6)] 



Light Type 



Optical Characteristic 



Power Source 

 Requirements 



Advantages 



Deficiencies 



IVIercury vapor arc 



Lines in violet, blue, green, 

 and Yellow, deficient else- 

 where. 



Current limited supply, 

 regulated power input 



Relatively high 

 efficiency (s 45-50 

 lumens/watt) 



Relatively long start time 

 (= 6 minutes) and restart 

 time (= 10 minutes). Poor 

 color rendition 



Thallium iodide- 

 doped mercury vapor 

 arc 



Green line of thallium 

 dominates the visible out- 

 put, mercury lines are 

 suppressed somewhat 



Current limited sup- 

 ply, regulated power 

 Input 



Very high efficiency 

 source of green light 

 (a 85-90 lumens/watt) 



Relatively long start time 

 (= 5 minutes) and restart 

 time {= 10 minutes) 



Dysprosium iodide 

 thallium iodide- 

 doped mercury vapor 

 arc 



Many lines and back- 

 ground level radiation 

 spread thruout the visible, 

 with the green thallium 

 line being the dominant 

 line 



Current limited sup- 

 ply regulated power 

 Input 



High efficiency source 

 of visible light (=80 

 lumens/watt) with 

 relatively good color 

 rendition 



Relatively long start time 

 {=A minutes) and restart 

 time (= 10 minutes) 



Incandescent tungsten 

 filament 



Continuous output through- 

 out the visible spectrum. 

 Increasing from the blue 

 to the red end of the 

 spectrum 



Regulated power 

 input 



Fast start time « 1 

 second) and good color 

 rendition 



Relatively low efficiency 

 (s24 lumens/watt) 



Xenon arc 



Continuous output through- 

 out the visible spectrum 



Current limited sup- 

 ply, regulated power 

 Input 



Fast start time (80% Relatively low efficiency 



output instantaneously) (= 22 lumens/watt), requires 



and Immediate restart, a high voltage pulse to start 

 excellent color rendition 



477 



