the solid rod into its housing and an arrange- 

 ment between claw and housing causes the 

 claw to close. A ball socket at the hull pene- 

 tration allows rotation through 45 degi'ees 

 on either side of the rod. At 1,000 feet deep 

 the external pressure causes the arm to re- 

 tract into the hull, but a nylon line and 

 pulley arrangement allows the operator to 

 extend the arm. In terms of freedom, it has 

 at least eight, including force feedback and 

 variable speed control. To collect samples a 

 cloth bag is reeled down on a string to an 

 appropriate position and, with the sample 

 inside, is reeled back up on the hull. In terms 

 of thru-hull waterproof integrity, both NEK- 

 TON BETA and GAMMA have been classed 

 by ABS; hence, one might properly assume 

 that this body of expertise found nothing 

 amiss. There is little else that can be said of 

 this approach except that at some point it is 

 depth-limited and that whoever conceived it 

 should have received an immediate raise in 

 salary! 



Claws 



Several types of claws or hands can be 

 seen in Figures 10.39 and 10.40. TURTLE has 

 a parallel type of jaw on the starboard ma- 

 nipulator and a Dorrance type of claw on the 

 other, beaver's claw is termed a hook type, 

 PISCES Ill's grasping claw is a circular 

 clamp, DS-4000's is an orange peel type, 

 DEEP view's is a clam shell, and NEKTON's 

 is a scissor type. The type of claw depends, 

 obviously, on the nature of the work. Orange 

 peel and clam shell varieties are best for 

 collection of soft sediment samples, the scis- 

 sors and Dorrance types of claws are good for 

 holding irregularly-shaped objects, and hook 

 types hold cylindrical objects well. 



Control 



Devices which the operator employs to ma- 

 neuver or control manipulators are either 

 portable or fixed, and each manipulator mo- 

 tion has its own button or switch on the 

 control device. Portable control boxes are 

 mandatory when the operator is moving 

 from one viewport to another and physically 

 cannot reach a fixed control. This is the 

 situation in ALVIN and an early version of 

 its manipulator control box is shown in Fig- 

 ure 10.41. Where panoramic visibility is 

 available, fixed controls are acceptable. Fig- 



ure 10.42 shows the fixed manipulator con- 

 trol panel on PC-14; each knob controls a 

 specific motion by dispatching hydraulic 

 fluid to the appropriate hydraulic line. 



Owing to the wide variety of tasks one can 

 envision for manipulators, it would be fruit- 

 less to make a blanket recommendation ap- 

 plicable to all vehicles. Experience, nonethe- 

 less, has provided some excellent general 

 guidelines to the manipulator designer 

 which are germane to all. In a 1966 article on 

 manipulators, Hunley and Houck (35) pre- 

 sented some lessons learned on TRIESTE I 

 and //. These lessons warrant due considera- 

 tion because the tuition was paid in the form 

 of lost manipulators and lost capability. The 

 following is extracted from their experi- 

 ences: 

 — Emergency jettisoning devices should 

 not allow inadvertent release of the ma- 

 nipulator. 

 — Some means to view the manipulator 

 when stored (to insure that it is, in fact, 

 stored) should be provided. 

 — Extended immersion in seawater calls 



for corrosion-resistant materials. 

 — Television cameras and lights on the 

 claw or forearm are invaluable aids for 

 fine positioning or inspection. 

 — External wire or cabling can be torn 



loose during tows. 

 — Internal leakage in joint actuators may 

 cause the manipulator to creep from its 

 last position. 



A more recent report by Pesch et al. (36) 

 discussed the test results of divers versus 

 manipulators in undersea work. While the 

 results are fairly predictable (the diver al- 

 ways won), several of the general conclusions 

 can be applied to increasing manipulator 

 performance, and others may serve to alert 

 the designer to conditions of which he might 

 not be aware. The following is extracted from 

 reference (36): 

 — Quite frequently the manipulator blocks 

 the work object from the operator's view. 

 — Requirements to align a tool perpendicu- 

 lar to a surface should be reduced be- 

 cause of optical distortion, empty field 

 conditions and the simple mechanical dif- 

 ficulty of alignment. 

 — The manipulator color should afford 

 moderate background contrast. A white 



531 



