The hydrostatic testing consisted of pressurizing a flange-mounted window 

 (Figure 5) until failure occurred. Since the window flange is open on one side to 

 the atmosphere, window fragments were ejected upon its failure (Figure 6). The 

 displacement of the window's low-pressure face during pressurization was measured 

 to ±0.001 inch by means of a wire that transmitted the displacement of the window 

 to a mechanical dial indicator over a pulley system without any mechanical ampli- 

 fication (Figures 5 and 7). To permit the attachment of a displacement indicator 

 wire to the center of the window's low-pressure face, a short acrylic rod with a 

 small transverse hole in one end was bonded to the window's surface with solvent- 

 type cement. The displacement of the window under hydrostatic pressure was read 

 directly from the dial indicator with a closed-circuit television system that permitted 

 the operators to be in a safe location during the ejection of the window from its 

 retaining flange when critical pressure was reached (Figures 8 and 9). 



As discussed in Appendix A, silicone grease was used as a pressure seal between 

 the window and flange. The grease was spread by hand on the contact area of the 

 low-pressure face and edge of the window. Sealing was completed when the window 

 was placed in the flange cavity, rotated in place and pushed inward against the 

 flange. This was done to distribute the grease uniformly over the area of contact 

 and also to eliminate any small air bubbles trapped between the window and flange. 

 This procedure proved to be adequate as it allowed no leakage of water to occur 

 between the window and the flange. Care was exercised to insure that both the 

 flange cavity and window were clean, since the flange was used for successive 

 testing and tended to retain small fragments of previously tested specimens. 



Since the ejection of windows In many cases fragmented them into very small 

 pieces, a reconstruction of the mechanism of material failure was usually impossible. 

 To provide data that would give an insight into the mechanism of failure, some of the 

 windows were pressurized only to a fraction of the window's critical pressure and then 

 removed for Inspection of their deformation and cracks (Appendix B). 



The explosive release of energy which accompanied window failure at higher 

 critical pressures was quite harmful to O-rings, bolts, and flanges. To decrease the 

 shock effects of this energy release, the cylindrical passage in the flange and the 

 adaptor flange was filled with water after the window was In place. At the moment 

 the window failed the water was forced through a l/2-inch-diameter restrictive 

 opening in the adaptor flange. This shock-damping method was sufficient to prevent 

 the breaking of the eight 1/4-Inch-diameter high-strength bolts connecting the 

 window flange and adaptor flange. 



